GPT-5.6

(openai.com)

716 points | by logickkk1 3 hours ago

106 comments

  • minimaxir 3 hours ago
    The developer's guide (https://developers.openai.com/api/docs/guides/latest-model) has some interesting semantic tips for using the model:

    > Intent understanding: GPT-5.6 can better infer the user’s underlying goal and intended level of work without you specifying every step. Continue to state important constraints, approval boundaries, and success criteria explicitly.

    > Original image detail: GPT-5.6 preserves the original dimensions of images sent with original or auto detail instead of resizing them to a patch budget or pixel-dimension limit.

    > Use shorter prompts: In internal evaluations, replacing long, explicit system prompts with minimal prompts improved scores by roughly 10–15%, while reducing total tokens by 41–66% and cost by 33–67%.

    > Avoid generic brevity instructions: GPT-5.6 is more sensitive than GPT-5.5 to instructions such as “Be concise,” “Keep it short,” or “Use minimal text.”

    > Control warmth: GPT-5.6 does not become meaningfully better when prompted to be broadly friendlier or more empathetic.

    • ravenstine 2 hours ago
      > Avoid generic brevity instructions

      That part is confusing because it's not like they provide an example of how default GPT-5.6 output compares with GPT-5.5 both with default output and prompted for brevity. Whenever I use such prompts, it's usually because I want the model to give me the gist in a few sentences. I'd be stunned if GPT-5.6 was that concise by default. I would think that could "break" a lot of things for developers who didn't know to make prompt changes after upgrading to 5.6. What if you were expecting GPT to be as wordy as it usually is? Then suddenly your output is not wordy enough?

      Smells like OpenAI trying its best to stave off financial armageddon for another few months. Then again, I'm not sure why they chose to waste so much output computation on verbal diarrhea all this time up to now.

      • tshaddox 18 minutes ago
        If you conceptualize this as “there is an appropriate amount of brevity for each situation” then it would be expected for a better model to use different amounts of brevity if it gets better at determining the appropriate amount.

        My view is that popular models by default output wildly excessive amounts of prose for nearly every use case, so if this changes in a new model that’s a pure win.

      • anticorporate 2 hours ago
        It seems like the way brevity instructions have changed is mis-aligned with how most people would expect to use them or are currently using them.

        Here's the example they give:

        > Instead of asking for the shortest possible answer, replace brevity instructions with prioritization:

        > Lead with the conclusion. Include the evidence needed to support it, any material caveat, and the next action. Omit secondary detail and repetition.

        > Keep all required facts, decisions, caveats, and next steps. Trim introductions, repetition, generic reassurance, and optional background first.

        Generally speaking, when I ask for a short answer, I want a short answer because I'm not really willing to read through a bunch of bullshit to get to a summary. Putting the onus back on me to assume what the model will return and write a longer prompt detailing exactly what information I want completely misses the point of why I'm asking for a short answer in the first place.

        • derefr 1 hour ago
          > Lead with conclusion.

          I would presume (perhaps falsely?) that an instruction like this would lead to the model presenting a conclusion not supported by the evidence, and potentially backtracking as it then tries to justify said conclusion.

          Yes, if deliberation happens, the model should figure out what it wants to say during that phase; but if you're using auto mode, the model is not going to be doing any deliberating half the time. In those cases, the output blathering is the model's only chance for deliberation. It "thinks as it talks", per se.

          Given that, I would advise a different approach: let it blather, but then get it to write you a conclusion at the end that the model can guarantee will obviate the need to read any of the blathering.

          I.e. advise the model to add an "executive summary" to the end of any non-trivial-in-length response. With some wording to carefully navigate the model between "the summary is itself too long" vs "the summary acts more like clickbait, leaving out necessary detail such that it requires actually reading the blather."

          Not sure exactly what that wording would look like. I imagine something like "write your postscript executive summary as if you were a senior CIA intelligence analyst summarizing ground-level reports into a daily digest for the Joint Chiefs of Staff. Take up as little of their time as possible, but ensure that any detail critical to decision-making is retained." (But that phrasing might only be useful if the model is delivering a certain type of response, and actively counter-productive otherwise. This kind of thing is delicate.)

          • Farmadupe 6 minutes ago
            Yes this is an extremely well known result for exactly the reason you guessed. It's not just abcktracking, asking an LLM to present a conclusion and then justify is also an excellent way to provoke hallucination as the model con concts "any justification that plausibly justifies the words it's already said".

            This is the actual reason why openai _invented_ reasoning models, to give them time/space to work out a solution, rather than having to magic a correct solution out of thin air from token 1.

            It's less important now that all models do reasoning, but it's still almost always better to make the output come out last rather than first.

          • prymitive 1 hour ago
            Oh the number of time LLM will, for example, be giving me the list of bugs it found in code, when I ask it for a review, just to decide there’s no big half way through explaining it.
          • skybrian 27 minutes ago
            I wonder if it would help to ask it to write a rough draft and then reorder it?
            • derefr 21 minutes ago
              It would (and does), yes; but this takes a lot more output tokens than asking for a summary would. The summary approach is only helpful insofar as it can be cheaper than using the thinking model. (You're basically tricking the instant model into thinking, which it can do, after a fashion.)

              But, unless your desired output is literally a document for others to read, at the point where you're having a model generate a full, lengthy output multiple times over with revisions, you may as well just turn off auto mode and have it always deliberate (i.e. choose the thinking model explicitly from the model selector.) Then it'll be as messy as it needs to be while deliberating, but give you exactly what you want as output.

              (And if your desired output is literally a document for others to read, that you want to interactively draft and polish, then (in the case of ChatGPT specifically) you should not only be explicitly forcing the "thinking" model, but also should be asking it to activate the "canvas" feature from the start. My understanding is that revising a canvas document involves the model emitting something like editing gestures, rather than simply re-streaming the updated chunks of text. This saves a lot of output tokens on large documents.)

          • radlad 1 hour ago
            I don't expect that would be the case. This is what's called BLUF or Bottom Line Up Front: https://en.wikipedia.org/wiki/BLUF_(communication)

            The model will still have read the entirety of the document before composing its response. And I believe that even in auto mode, there are thinking tokens behind the scenes.

          • cma 1 hour ago
            Why would auto mode turn off thinking?
            • derefr 1 hour ago
              The "auto" mode is (AFAICT) a per-conversation-turn router. (Presumably via a preliminary pass through a very fast tiny model that spits out an number for how challenging it thinks the next response might be to compute.)

              On high-challenge turns, the auto mode routes to the "thinking" model. But on low-challenge turns, it routes to the "instant" model.

              And the "instant" model, by design, has no capacity for deliberation. (If it did, it couldn't guarantee that its responses would begin streaming "instantly.")

        • spathi_fwiffo 2 hours ago
          Replace 2 word instruction ('be concise') with a 38 word instruction.

          Human can no longer be concise when asking for a few sentences instead of 20 paragraphs of BS they don't want to read when all they want is a summary to verify the general direction of the prompt-work before digging into the details.

          such progress!

          • osigurdson 2 hours ago
            I don't know how intentional it is / was, but LLMs in general just love to hear themselves talk!
            • arjie 2 hours ago
              They do, and I want to encourage them to do so because they think through talking. What I don’t want to do is spend time reading all that.

              We will probably just get reader-side affordances for this like auto-folded justification and introduction sections and so on.

              Doubtless some chat interface will add this the way they’ve added reasoning folding.

              • bcrosby95 2 hours ago
                Thinking models think through talking, don't reveal that talking, then answer by again thinking through talking. It's kinda funny in a way.
            • mr_toad 33 minutes ago
              > LLMs in general just love to hear themselves talk!

              Because that’s what’s in the training set. Reticent humans don’t have blogs.

              • flir 7 minutes ago
                Interesting idea. I think they're getting more wordy over time, personally, so I think it's more to do with the training than the raw data.
            • jimbokun 2 hours ago
              Is it just a coincidence that the companies creating them charge by the token?
              • pizzafeelsright 2 hours ago
                The aligned incentive appears to be realigning in favor of the corporation.

                Pray they do not realign them further.

                There are times I require single word answers. I will use whatever model responds as I desire and at this point those models are just a few.

                • minimaxir 1 hour ago
                  The cost-per-task benchmarks align incentives toward more efficient output and those are the ones gaining steam.
        • Romario77 1 hour ago
          I think instead of "be concise" you could tell it how long the answer should be. I.e. give the answer in one paragraph. Or in 10 lines max.

          At least before it would listen to instructions like this.

          • isityettime 1 hour ago
            > At least before it would listen to instructions like this.

            Would it actually follow them? IME LLMs are incapable of estimating the length of their own output, the total length of the current context, etc. They just make stuff up unless they have external tools that can inspect those things for them.

      • ignoramous 1 hour ago
        [flagged]
    • egorfine 1 hour ago
      > Intent understanding

      This will totally make it brain damaged over a certain tasks. Sort of like the same brain damage that prompted OpenAI project managers to destroy ChatGPT.app today.

      • oezi 6 minutes ago
        Can you elaborate?
    • avaer 1 hour ago
      I'm impressed. It feels like a faster Fable (probably due to the more efficient token usage). It performs roughly the same job, just with 4x less steps (gamedev).

      Remains to be seen how the "shorter prompts" advice translates to homogeneity/collapse though.

    • mlmonkey 2 hours ago
      > Avoid generic brevity instructions: GPT-5.6 is more sensitive than GPT-5.5 to instructions such as “Be concise,” “Keep it short,” or “Use minimal text.”

      What about my favorite, "no yapping"?

      • astrange 24 minutes ago
        It might need the longer answer to think about the question, so one approach would be to ask it normally and then ask it to repeat itself shorter.
    • swatcoder 1 hour ago
      > can better infer the user’s underlying goal and intended level of work

      This is a trap.

      It's the optimistic fallacy that poisons all "consumer scale" machine learning products and what's going to effectively ruin these models as they keep chasing it in the same way that web queries were ruined, social media feeds were ruined, and media recommenders were ruined.

      For the vendor, optimizing metrics across their whole user base, they always see positive technological progress as their system gets better at making assumptions and accumulating user engagement scores in aggregate. But for the individual user, most of which has some weird tail intent/interest and some of whom have many weird tail intent/interests, the experience quietly but catastrophically degrades. Output/results become more generic, more divergent with the underspecified "weird tail" intent, and more stubbornly hard to ever wrangle towards that "weird tail" altogether.

      We've been watching this cycle happen for 20 years now and it's proving hard for anybody to escape because it works so well for the trillion dollar company driving it forward. But while each step might feel ergonomic and welcome to individual users, there's a frog boiling enshitification at play.

      In pursuit of output quality and capability (rather than simply the vendor's user count), what we need rather than "makes better guesses" is "presses for more clarity", even where it feels kind of annoying.

      Even among human professionals, one of the first hurdles of breaking out of junior tier work is gaining the confidence to press your colleagues and clients to be more specific in their thoughts and expressions despite their desire to have you do it all for them. But they're often coming to you with incomplete, muddy, and conflicting ideas for which there is no safe and correct assumption that you might just run with, and it's your expertise (i.e. relevant "intelligence") that's critical to bringing attention to that. To achieve professional progression, you need to learn to do that and to not just optimize appeasing the ambiguous client/colleague today in exchange for mutual expense tomorrow. To avoid enshitification, which is probably not possible, we need these models to be learning that too.

    • artisin 2 hours ago
      Control warmth[1]

      > GPT-5.6 does not become meaningfully better when prompted to be broadly friendlier or more empathetic. Instead of generic instructions such as “Be friendly and warm,” use concrete guidance: > Be direct and tactful. Acknowledge friction specifically when relevant. Avoid canned reassurance and unnecessary sign-offs.

      Soo basically, my new 5.6 custom instructions: Be Jeeves and eliminate all friction from my life through immense processing power. Acknowledge friction specifically when relevant. Avoid canned reassurance and unnecessary sign-offs.

      [1] https://developers.openai.com/api/docs/guides/latest-model#c...

    • stillpointlab 1 hour ago
      > Avoid generic brevity instructions: GPT-5.6 is more sensitive than GPT-5.5 to instructions such as “Be concise,” “Keep it short,” or “Use minimal text.”

      I used to go to a barber and if you said "cut it short", he cut it really short.

    • epihelix 2 hours ago
      > Use shorter prompts: In internal evaluations, replacing long, explicit system prompts with minimal prompts improved scores by roughly 10–15%, while reducing total tokens by 41–66% and cost by 33–67%.

      When has this ever not been the case? I don't think this is a GPT 5.6 specialty!

      • adam_arthur 2 hours ago
        Information density of the prompt is the most important factor in my experience.

        And interestingly, LLMs seem particularly bad at writing prompts for other LLMs for this reason (you can guide them to be more dense, just speaking by default).

        Conciseness is usually a byproduct of information density though.

        • docjay 20 minutes ago
          Lexical-priming->semantic-space-constraint;specialized-lexis+=sharp distributional-signature;∴ tight concept-cluster; generic-lexis->diffuse-activation, broad candidate-set;Attention-heads key/query-match domain-tokens;"Hamiltonian"->{operator,eigenstate,quantum,energy}->register+domain locked;Net:constrained-decoding,vocab=soft-prior over output-distribution; register-matching;#taskdef=decompress->continue
        • astrange 23 minutes ago
          > LLMs seem particularly bad at writing prompts for other LLMs for this reason

          Claude is terrible at this! Probably for the same reason that its writing style in prose is so annoying and full of claudisms.

      • daemonologist 1 hour ago
        There was a fad a while back of building insanely long prompts - tens of thousands of tokens - including having models write prompts for themselves. I always thought it was counterproductive, especially if you're going to use the prompt more than a couple of times. (That said, the e.g. Claude Code system prompt is insanely long, so if you genuinely have a lot of information to provide maybe it's beneficial. Like, shorter is better, but you don't want to be under-specified.)
      • CuriouslyC 1 hour ago
        For Gemini 2.5 and ~GPT5.0-5.1, longer prompts with lots of explicit instructions and examples produced better conformance. Seems like heavily second guessing the models started to get counter productive around the end of last year.
    • postalcoder 2 hours ago
      > Avoid generic brevity instructions: GPT-5.6 is more sensitive than GPT-5.5 to instructions such as “Be concise,” “Keep it short,” or “Use minimal text.”

      RIP Caveman skill. Six month good. Now skill dead.

      • gershy 43 minutes ago
        Caveman speak make compression not brevity
      • delichon 2 hours ago
        A Yoda skill, is there?
        • froh 44 minutes ago
          > Yoda skill, there is?

          ftfy

    • elAhmo 2 hours ago
      > Use shorter prompts: In internal evaluations, replacing long, explicit system prompts with minimal prompts improved scores by roughly 10–15%, while reducing total tokens by 41–66% and cost by 33–67%.

      A shorter prompt results in half as much tokens spend? I find this very hard to believe.

      • bulder 2 hours ago
        If it's anywhere close to the same universe as smaller models in its behavior, a lot of time in "thinking" mode is spent on reiterating on any constraints given in a prompt. So the more constraints you give it, the more tokens it will spend going "Hold on, the prompt said I have to dot my i's and cross my t's. Let me go through my work to check that all the i's are dotted."
      • zeven7 2 hours ago
        Maybe Codex has the same problem I sometimes have focusing while reading and has to reread the same sentence over and over again.
    • firemelt 2 hours ago
      do we have similar guidance or page from anthropic for claude?
    • cromka 1 hour ago
      [flagged]
  • simonw 1 hour ago
    Here are 18 pelicans - six each for Luna, Terra and Sol at the six different reasoning effort levels (plus the price to generate each one): https://static.simonwillison.net/static/2026/gpt-5.6-pelican...

    Or if you want to see some in 3D, OpenAI featured a pelican riding a tricycle, bicycle, pony and another pelican in their livestream this morning: https://www.youtube.com/live/Wq45rvPGNHs?t=1070s

    • pupppet 39 minutes ago
      Time to dump this test. Probably not a coincidence every version has the same rolling green hills, gradient blue sky, sun in the corner, etc.
      • fastball 19 minutes ago
        On the one hand: yes, pelicans on bikes are definitely in the training set at this point.

        On the other hand: the test is clearly not saturated, given that you can see a clear difference in output at the various reasoning levels / model versions.

    • Night_Thastus 38 minutes ago
      I think the 'pelican test' is becoming useless. It's been around long enough that now I'm sure good examples are in the training data, and hell they might even do some hand tuning to make it do a decent job since they know people will ask about it.

      But either way, with no real way to visualize the result of the text it starts with - it will always be stabbing in the dark. It can't understand conceptually what any of it should look like and then refine the SVG to improve it gradually. It just throws darts at a wall and hopes it comes out alright.

      • skybrian 34 minutes ago
        I think it's still useful in a "hello world" sort of way. It means you actually tried the new model.
    • irthomasthomas 26 minutes ago
      Cool. I still find these a useful visualization of some the qualities of llms. Even if they did train for [animal] on [vehicle] svg, it's still nice to see at a glance how the different models and reasoning levels perform. Lunar misses part of the frame, except on max reasoning. While most of the others have a mostly correct bike at all reasoning levels.

      I once used something like karpathy's auto-scientist to mutate the prompts and rank them with a vison model. Some of the winners where pretty neat. I think they have a lot more style than the gpt-5.6 ones. https://xcancel.com/xundecidability/status/20449185674144196...

    • cma256 59 minutes ago
      Is the direction of the pelicans encoded in your prompt? Curious why they are all left to right with the exception of terra xhigh.
      • sbayg 44 minutes ago
        History moves left to right.
    • JeremyHerrman 14 minutes ago
      people are saying this is benchmark is saturated but all of these have occlusion issues, even sol max.

      A skilled human artist wouldn't have both legs in front of the bike, or a single straight line representing both leg's crank arms.

      • MichaelZuo 11 minutes ago
        Yeah it makes no sense at all to dismiss the test, when even the very best examples are noticeably below what a skilled teenager could produce.

        Dead internet theory? Semi-random parroting by real people? Or something else.

    • ashu1461 23 minutes ago
      They said in the AI community, a pelican riding a bicycle is a good test to measure effectiveness of the model, wondering if they were referring to you, or is it really a standard in the AI community ?

      Also would be good to have a tool where users can select models and instantly see each model's generated pelicans. That will make it easy to compare the output of different models.

      • jjice 22 minutes ago
        Simon did start the pelicans on bicycles as an SVG, but I think it's more of a fun goofy thing to see how the model performs at. I don't think it has a direct correlation to a model's performance though.
    • miohtama 56 minutes ago
      At what stages will models start to internally reflect the drawn SVG and automatically fix their own mistakes?

      I assume multimodal models can do it already do it today if constantly asked "make it better"

      • simonw 27 minutes ago
        I haven't tried this in a few months, but last time I tried a loop that rendered the pelican and asked for improvements the results were actually quite disappointing. Be interesting to try that again against GPT-5.6 at Claude Fable 5 though.
    • aappleby 52 minutes ago
      Surely "how to draw a SVG pelican on a bike" has made it into the training data by now ...
      • astrange 39 minutes ago
        If that was the case in a non-trivial way you'd see mode collapse, but you don't, they come out differently.
        • tlamponi 34 minutes ago
          It's because all the frequent comment that this pelican is in the trainings set now also got into the trainings set and models adapt. /joking (I hope)
      • huseyinkeles 47 minutes ago
        Surely this comment is literally on every new model release post.
        • simonw 28 minutes ago
          It's part of the pelican tradition at this point.
      • conception 47 minutes ago
        And yet... only one is any good.
    • zniturah 9 minutes ago
      max effort sol clearly over-engineered
    • siva7 1 hour ago
      Ok, I'll never use max effort again on OAI models..
      • isoprophlex 35 minutes ago
        Is that... an x-rated, censored pelican?
    • throwaw12 1 hour ago
      something is wrong with Terra model series, most pelicans, except Max, looks bad
      • CryZe 20 minutes ago
        Seems to match the pareto frontier on Artificial Analysis as well. Terra is nowhere on it.
    • yokoprime 1 hour ago
      Thank you Simon! Luna is surprisingly decent across all reasoning levels.
      • vecter 1 hour ago
        I think all of Luna's are bad. The only decent one is sol @ xhigh. Even sol @ max is weird. Sol @ high and @ medium are ok, and every other single one across every model is bad.
        • hn_throwaway_99 51 minutes ago
          Strong disagree, but to each their own. For sol I really like how only medium uses the wings on the handlebars to ride the bike. For all the other sols the pelican evolved a new set of arms separate from the wings.
    • ismailmaj 46 minutes ago
      Apparently plus users do not have access to Sol, so I'm really worried about the ugly Terra Pelican.
    • algoth1 56 minutes ago
      gpt-5.6-sol Max pelican didn’t skip neck day
    • firebot 56 minutes ago
      AI really sucks at bicycles...
      • rsoto2 14 minutes ago
        LLMs really suck*
    • emehex 1 hour ago
      gpt-5.6-sol x XHIGH is my favourite
    • btown 57 minutes ago
      You're famous!
    • rsoto2 14 minutes ago
      this looks like the same shit from 4 years ago. give it up.
    • hirvi74 58 minutes ago
      I like Terra High the best. That pelican is utterly yoked.
    • alecco 28 minutes ago
      Why is this AI influencer slop comment always on top? There are plenty of good comments with actual signal buried.

      One has useful information on how to prompt. Another comment shows it tops the ARC AGI v3 significantly.

      But no, the top comment has to be the repetitive and meaningless SVG test.

  • meetpateltech 2 hours ago
    GPT-5.6 Sol sets a new SOTA on ARC-AGI-3: 7.8%

    Sol is the first verified frontier model to ever beat an ARC-AGI-3 game

    https://arcprize.org/results/openai-gpt-5-6

    • 10xDev 2 hours ago
      Seeing the dramatic differences in scores just going from high to xhigh is just another demonstration of the bitter lesson: Just keep scaling search and learning. We are probably going to need a lot more GPUs.
      • energy123 3 minutes ago
        > Dramatic difference

        Isn't this just the difference between getting 0 right and getting 1 right?

      • bevekspldnw 1 hour ago
        These aren’t raw base models they are the result of a ton of RLHF and various adjustments.

        Bitter lesson wildly overstated in this context.

        • froh 42 minutes ago
          rlhf = reinforcement learning from human feedback

          (had to look it up)

        • cma 10 minutes ago
          The scaling with reasoning models is more and more with things like verifiable rewards (coding and math), in line with bitter lesson and also Sutton invented lots of modern RL.
        • fastball 21 minutes ago
          More RLHF is in fact scaling.
      • altcognito 2 hours ago
        While I think this is true, remember as we get more efficient we just decide to scale even bigger. So more GPUs, and more efficient.

        I agree with the sibling comment, effiency is probably the more important component at this point. We are hitting not just a practical engineering roadblock for scaling with current technology, I think we have definitely hit a financial and logistical roadblock for up scaling with the number of GPUs (on an immediate basis)

      • Salgat 1 hour ago
        Kind of refreshing though that the "throw more processing at it" scaling we saw in the 90s has returned in a different way. For a while we were really bottlenecked in our advances by relatively low levels of parallelism (most software used by your average user doesn't scale cleanly with more than a few threads).
      • vatsachak 2 hours ago
        I mean, theoretically you can solve every finitary problem with a brute force solution...

        Richard Sutton specifically states that the search has to be smart. We know that the brain uses recurrent connections and is shallow. I think a lot more money has to go into architecture. Feed Forward transformers can only scale so far

      • dyauspitr 21 minutes ago
        And dozens of data centers in every state so tokens are dirt cheap.
      • emp17344 1 hour ago
        This isn’t really how it works anymore. Agents rely heavily on tool use and the agentic harness to perform tasks. Pre-training is no longer very effective.
        • HDThoreaun 1 hour ago
          I thought models werent allowed tools on arc-agi?
      • Razengan 2 hours ago
        > We are probably going to need a lot more GPUs.

        Or a breakthrough in algorithms etc.

        The human brain, heck all bio brains, are proof that you don't need a lot of power or size for intelligence.

        • ryandvm 1 hour ago
          The human brain has 80 billion neurons and a 100 trillion synapses. I think you're underselling the processing power of that warm chunk of meat.

          The real message of the last 15 years has actually been the opposite: if you throw enough processing power at it, intelligence emerges.

        • altcognito 1 hour ago
          20 watts for inference AND training!
        • aeyes 1 hour ago
          For intelligence, I expect the next breakthrough to be colocation of memory and compute in the same chip. And we'll need much more of this memory, probably a few petabytes.
    • simianwords 2 hours ago
      Very interesting. My prediction is that Mythos would outperform Sol.

      Also what does this tell about Yann LeCuns whole world model theory? Bro has been going on and on about it. He has made multiple wrong predictions on the trajectory of LLMs.

      At some point his claim should be fully falsified no?

      • osti 2 hours ago
        Mythos probably wouldn't, otherwise they'd have included it in their release. Next version of Mythos probably will though.

        And yeah.. Reality has not been kind to LeCun.

        • vatsachak 2 hours ago
          Are you joking? They spend billions of dollars training LLMs to get a 7.8% on arc agi 3 whereas DINO models are near sota in image classification, provide meaningful embeddings to the point where image segmentation is just PCA. The spend on DINO cannot be more than five million (correct me if I'm wrong)

          JEPA is just getting started

          • Tenoke 1 hour ago
            His main anti-LLM predictions have been consistently either wrong or misleading.

            There's many ways to skin a cat so you can probably do something with a JEPA approach as well, but I doubt he actually catches up to having agents on the level of where Anthropic/OpenAI will be at any point.

            • onlyrealcuzzo 1 hour ago
              His main LLM predictions have almost nothing to do with Arc AGI...

              What exactly was he dead wrong about that is proven by any of this?

              GPT getting better has absolutely nothing to do with completely disproving anything LeCun has been saying.

              He never said LLMs couldn't get better. He never said they couldn't score 7.6% on Arc AGI 3.

              He's merely said they don't think, and you probably want something that actually thinks if you want a model that can be trained cheaply on a small amount of data and provide a ton of value.

              Spending $5B to train a model that scores better than an older model does not disprove any of that in any way.

              • Tenoke 1 hour ago
                >He's merely said they don't think

                He said years ago even 'GPT 5000' couldnt do things that they ended up doing fine a month later, let alone by 5000. His later predictions are just moving that goal post including towards them not being able to do more general, harder problems of which Arc AGI is a counter-example.

                • onlyrealcuzzo 1 hour ago
                  > He said even 'GPT 5000' couldnt do things that they could do a month later, let alone by 5000.

                  What things specifically and when?

                  • Tenoke 1 hour ago
                    https://youtube.com/shorts/zQTt8TkcyfU?is=09r7XDqz2w6-Pygu

                    You probably wont like the edit but I dont have the timestamp of the original on hand, you can find it.

                    • onlyrealcuzzo 1 hour ago
                      That does not at all look cherry picked or taken out of context...

                      LeCun's ideas cannot be reduced to a 6 second clip...

                      You're missing the forrest for the trees, taking a singular example of a problem and thinking that if an LLM can solve the singular example it completely disproves LeCun is comical...

                      • Tenoke 1 hour ago
                        You can watch the whole Lex Friedman interview, it's on youtube. It's not out of context at all. He goes on about how LLMs will never be able to do things that they do trivially. And he has just doubled down for years.

                        Ive read and watched more of his interviews and lectures it seems, it feels like you just have a rosier idea of his views than the views he repeatedly presents.

              • cma 7 minutes ago
                > What exactly was he dead wrong about that is proven by any of this?

                He said as you need more and more tokens models will fall apart because each additional token is a chance for a mistake and they will just exponentially fall apart. But in practice models have learned to identify and self-correct mistakes and if you look at the graphs more inference reasoning tokens almost always give far better accuracy.

          • redactsureAI 2 hours ago
            DINO is a transformer model?
          • esafak 2 hours ago
            ASI is going to be here by the time Lecun gets started.
      • Maxatar 1 hour ago
        Falsifying Yann Lecun isn't exactly a priority for anyone seriously working in this space.
      • bevekspldnw 1 hour ago
        “Bro” spent most of his career in the wilderness because everybody thought ML/NN/etc were a dead end.

        I’d not wager against him having at one one more break though architecture before he retires.

  • rsanek 5 minutes ago
    Based on the Intelligence vs. Cost graph, not clear to me why anyone would use Terra? Luna looks quite interesting though, happy to see OpenAI still serving the more budget-oriented side of the market (seems like Anthropic and Google have lost interest there).

    https://artificialanalysis.ai/articles/gpt-5-6-has-landed

  • Syntaf 3 hours ago
    Ok long time Claude Code user here; lately I've started to realize there's other great models out there I should be trying, but I'm hesitant to leave Claude Code behind for something new.

    What's the consensus today on codex vs claude code, does it really matter anymore?

    • nilkn 2 hours ago
      Codex has arguably been better than Claude Code for months now, but it's flown under the radar because it just didn't capture the same viral marketing effect and OpenAI in general has had more optics / PR issues than Anthropic amongst the online developer crowd. I use the word "better" not in the sense that the underlying GPT models are fundamentally smarter or more intelligent, but rather that as a product Codex is just simpler, cheaper, and abundantly reliable and low-drama.
      • hk__2 2 hours ago
        I’d argue the opposite. I’ve switched back and forth from one to the other and Opus/Fable has been constantly better than any GPT in my daily work. It’s a bit slower but it does the things right, with as little code as possible, some comments where needed. Codex is faster but you always have to correct it because it got something wrong; it writes tons of code ("let me add a small helper") with obvious comments.
        • ljm 2 hours ago
          Purely anecdotally the one persistent issue I have with LLMs writing code is that they are absolutely paranoid and add a load of indirection and defensive crap and even if you prompt to avoid that it will often require manual steering to remove the cruft.
          • pdantix 5 minutes ago
            recent gpts are horrendous for this, whereas recent claudes have a tic where they incessantly add useless comments referring to previous changes and will use multiple single-line comments instead of a standard multi-line docblock.
          • hk__2 40 minutes ago
            I’ve experienced this with GPT but not with Opus/Fable.
          • adastra22 1 hour ago
            I have not noticed this with Opus 4.6+. The result is usually not too far from what I would have written myself.
        • cevn 2 hours ago
          Sounds like you are talking past each other. GP is saying the harness of codex is higher quality, which I can believe, even if the models are not as good as Opus/Fable.
          • phoghed 5 minutes ago
            GPT-5.5 is as good though, at least according to my personal experience and DeepSWE
        • nilkn 2 hours ago
          I'm not sure how meaningful this is. Fable only just recently become more broadly available, and GPT-5.6 is launching broadly today.
        • dbbk 2 hours ago
          I really love the Opus/Fable models but I'm honestly sick to death of the buggy product. The CLI always has some weird issue. Right now it doesn't even output messages before tool calls, it just swallows them and they disappear.

          I don't like OpenAI as a company, but they appear to have QA, and that is probably enough to get me to switch.

          • walthamstow 2 hours ago
            There was an issue on Claude Code the other day where it would only wait 60 seconds when it had asked a set of questions, then if it didn't get a response from the user it would just continue however it thought was best. Completely unusable. It took them nearly 48 hours to merge a fix.
          • pqdbr 34 minutes ago
            Glad I’m not the only one noticing this. It’s maddening.
        • behnamoh 1 hour ago
          > Codex is faster but you always have to correct it because it got something wrong

          this has been my experience with Codex as well, and I have to fix its mistakes every single time. But recently, I literally threw away three hours of work because it kept adding hundreds of lines to my code base. When I restarted the entire work using Fable and Opus, it was like night and day.

      • jrflo 2 hours ago
        Agreed. GPT 5.5 will come up with more straightforward solutions with far fewer tokens than Claude. Also, the usage limits are much more generous for Codex than Claude Code for the same monthly plan.
        • mattmanser 2 hours ago
          Last time I used Codex it would make loads of assumptions, often quite big ones, without asking.

          Did they fix that, as that for me was what actually made codex worse.

          • arcanemachiner 1 hour ago
            I find that I have to tell GPT and Claude to keep asking me questions, or they will just fill in the gaps themselves (wrongly).
          • nico1207 1 hour ago
            Did you use plan mode?
      • nolok 2 hours ago
        I really want a good Claude Design competitor in Codex, it's hard to use the others after getting used to it and yet I find anthropic's model to have a much worse understanding of what looks good or not than OpenAI or Google models.
      • dyauspitr 2 minutes ago
        Honestly it’s the usage limits that are so generous that makes codex worth it even if it may not be exactly as powerful as Claude. The peace of mind that you can try a lot of things and make huge refactors and run extensive redundant tests without running out of tokens just makes the whole thing a much better experience. I tried coding with Deepseek and it was pretty terrible so the only reason codex works is because its abilities are close to or on par with Claude.
      • Certhas 2 hours ago
        That's a strange statement... It's been true for a while now that OpenAI has had much more generous limits than Anthropic on their subscription plans. And with the Fable ban/guardrails disaster, there has been a lot of frustration from people in these comment sections. And Anthropic fucked up Claude Code pretty badly for a couple of weeks during the 4.6/4.7/4.8 transition, which again was widely publicized. And they got a lot of flack over not allowing other harnesses anymore. And ChatGPT got some pretty viral wins on model intelligence when they cracked the high profile Erdos problem.

        If anything the online optics have been bad for Anthropic for the last half year. OpenAI doesn't have optics issues, from my point of view they simply have the issue that they are the least trustworthy player at the frontier. The way they pivoted from their original mission is truly breathtaking, especially coming in gloatingly to take the government contract when Anthropic got kicked out for insisting the government does not use their systems for mass surveillance or autonomous weapons systems. You understand what that means, right? OpenAI models are now actively used/developed for mass surveilance and/or autonomous weapons systems.

        I know there are plenty here who seem to value their own ability to use these models cheaply above all other considerations. Then OpenAI is a great choice, and much less restrictive than Anthropic. But their problem is not on the optics. It's on the substance.

      • Trasmatta 1 hour ago
        Switched to Codex last week, and I'm already MUCH happier than I have been with Claude Code. Which surprised me.
      • ramijames 1 hour ago
        What about cost?
      • dboreham 2 hours ago
        Nudged by this thread, I've decided to switch from Claude to Codex for a bit to see what happens. But...I immediately became lost in their marketing vortex of confusion on plans and pricing. Anyone care to tell me which plan I should be using? On the other side I use the $100 Claude Code plan. We actually have a "Business" ChatGPT subscription already, which seems to be $50/mo/seat. OpenAI's web site offers a set of individual subscriptions (for parity with CC presumably) which I suspect weren't available when we signed up for ChatGPT. I think that in turn happened due to some web site feature it didn't allow for free users (uploading PDFs, something like that). Perhaps I should switch from that business account to an individual subscription for Codex?
        • gruntled-worker 1 hour ago
          Test-drive it with an individual Pro account (5x or 20x) for a month. Download the Codex CLI client from https://github.com/openai/codex and auth it in the browser via the URL it provides. Set the model to 5.6-Sol and effort to max.
    • postalcoder 3 hours ago
      There is so much less drama involved with the Codex world. You don't realize how oppressive CC is until you've escaped it. Outages, weird restrictions, degradation, accelerated usage, etc etc etc.
      • timcobb 2 hours ago
        Totally. My experience as well. After some time with codex you're like come on Claude can you just stfu! Haha. I now almost always instruct Claude with specific length requirements when I ask questions. Otherwise, it just blathers and blathers in the most annoying of ways. "Oppressive" is spot on in my opinion
      • jakswa 2 hours ago
        I'll agree and expand on "weird restrictions" -- I used to check the claude usage graphs multiple times a day to see where I'm at on my weekly budget. With gpt 5.5 I don't think I'm working differently but haven't felt the need to check anything because I think I've hit my limit... once? on some egregious edge case scenario iirc
      • Cider9986 2 hours ago
        Even less drama with open models like GLM.
      • Amir6 2 hours ago
        Let alone getting banned out right with no reason, zero updates after weeks, and not even being able to download your chat history (despite the feature being available (I assume they vibe coded it and it does not work!). My story below;

        https://news.ycombinator.com/item?id=48597861

    • wwind123 2 hours ago
      I've been using Claude Code, Codex, Gemini (now Antigravity) at the same time for half year now, ever since I dipped my toe into agentic coding. I'd say in general Claude Code and Codex are equally powerful, Gemini is lagging behind.

      One thing I appreciate with Codex is, OpenAI nowadays sometimes just gives you quota resets you can bank, so when you use up weekly quota before the week ends, you could just reset the quota, to continue using Codex. I've been much less anxious about Codex quota because of this perk. I just used one reset in the bank yesterday, and still have 3 resets left. Whereas with Claude, when you've used 95% quota 3 days before the week ends, you'd be much more anxious.

      On the other hand, Claude Code's /remote-control mechanism is extremely helpful when I am running it in the cloud and wants to monitor it or control it on my phone. Codex currently doesn't support this kind of usage. Codex only allows you to use your phone to connect to a session on your desktop, not in the cloud.

      • EMM_386 2 hours ago
        Yes - Anthropic badly needs this same "here's a reset, use it when you want".

        It's vastly better this way. Sure, it may impact the bottom line but it's a huge customer satisfaction win.

        When Anthropic randomly resets me and I've only used 2%, that's worthless. When OpenAI tells me I have 3 resets available to use whenever I want - it's wonderful.

      • WhitneyLand 2 hours ago
        Codex is supported well on iPhone/iPad, it’s inside the ChatGPT app.

        It’s amazing how much work you can get done on your phone now, especially if you already have a design mapped out in your head.

        • ghostpepper 2 hours ago
          I have used claude and codex extensively but only from their CLI app (heavily sandboxed using rootless podman, network filtering, etc), so I don't really know what I'm missing with the GUI apps.

          One killer feature that Claude has, and AFAIK Codex still lacks, is the ability to start a session in the terminal and then hand it off (actually just remotely control it), from the iOS app.

          Last time I tried Codex on iOS it required a ton of set up to link a github project etc. The way claude lets me remote into a session I've already started on my actual machine is much better IMHO.

          • akmarinov 2 hours ago
            They’ve addressed that. Codex in the ChatGPT app on iOS is way better than Claude Code now.

            You sign in the Codex app on your Mac same on iOS and are able to completely control your sessions - fork, side chats, plugins - everything.

            It’s really great i often work through it. And you can connect any number of Codex instances on any number of macs and then manage them all through the iOS app.

            • ghostpepper 1 hour ago
              maybe I'm misunderstanding but I don't want to sign into the app on a mac - I want to run the CLI on a headless linux server and control the sessions from my iPhone. Does Codex allow that now?
              • frostterra 43 minutes ago
                You can do it too. There are two ways:

                1. Run `codex remote-control --help` directly on your Linux server. 2. From the desktop app, connect to your Linux box, start Codex there, and make it remotely controllable.

                Either approach will get you set up.

            • wwind123 1 hour ago
              Hmm, I don't have a desktop computer. I prefer my laptop be used for other purposes, and can sleep when not in use, instead of running a coding agent 24x7. That's why I prefer running coding agents in the cloud.
              • akmarinov 1 hour ago
                You can do that too. Start working locally, then just do /handoff to transfer your session to the cloud and then work through the Codex app on your phone.
      • _superposition_ 2 hours ago
        I've been using codex app server. Works great.

        https://learn.chatgpt.com/docs/app-server

        • wwind123 1 hour ago
          Hmm, thanks. Didn't know about this. But looks like a bunch of hassle to set it up?
          • _superposition_ 1 hour ago
            Finding the right docs/flags took longer than anything else. 15 mins from zero to productive on my phone.
            • _superposition_ 1 hour ago
              One thing I noticed you can't use codex installed via npm, but it will tell you that. Ymmv I'm a pretty simple user and do things in small manageable chunks. I try keep context < 10% before I fire off a task so I don't use many skills etc.
          • frostterra 41 minutes ago
            if you don't want to hassle with it, use desktop app, you either can make it remote controllable or you can control other devboxes.
    • davidhs 1 hour ago
      I recommend trying Codex too. In fact, I recommend running them side-by-side if you have the budget, e.g. have both independently plan the same feature or implement in a different worktree, or have them critique each other's work.

      I personally find GPT-5.5 to be a better programmer than Opus 4.8, it is extremely thorough, but I don't like the code it generates ("austere"), and find Opus 4.8 to write more "human friendly" code. The programming comments GPT-5.5 makes is pretty awful where-as Opus 4.8 is good. I feel like Opus 4.8 is better at grasping my intention than GPT-5.5, and honestly find GPT-5.5 to be kind of "autistic". I do prefer the language (not the writing) of GPT-5.5, as I find the philosophical flowery language of Opus 4.8 kind of annoying.

      I have only managed to try Fable 5 a little bit, which feels like a much more generally smarter version of Opus 4.8, that is much better a programming and grasping your intention, and I think even the intention of your code, and is _really_ good at spotting bugs or problems with logic in your code. It feels wicked smart but is extemely expensive. It feels smart in the sense like it has a "bigger brain" and is much more sensitive to subtleties/details.

      These are different "brains", have different "personalities", etc. I think the best thing is to develop a feeling for it yourself.

      • novaRom 19 minutes ago
        I haven't tried Codex yet, but I for my tasks GPT-5.5 may correctly point to a proper direction but its code feels a bit weird. Opus 4.8 is way better in coding, and actually it's the only one who could catch very very sophisticated bug in a large codebase (I tried different models including GPT-5.5 and DeepSeek). Interestingly Gemma 4 under opencode running locally performs not bad at all, it's far yet from DeepSeek level, but it manages to understand tools quite well, and code quality is pretty good. So, for simple coding projects I can say local models already won. It's amazing how smart open models of desktop size have become today. I mean it's quite plausible to manage small codebase today relying on only open tools and local models, you don't need any subscription to produce high quality code, but yes I assume you already experienced and know what you're doing :)
    • pkulak 2 hours ago
      I can't tell the difference between Fable and GPT 5.5. I tried Fable while it was in trial $20 mode, used up my whole quota, and it was great, but as soon as I went back to GPT 5.5, everything was the same.

      But what I love about Openai is that they still let you hook OTHER harnesses up to a subscription. My Pi setup has been built up for a few months now into exactly what I want and moving over to CC or even Codex is really annoying.

      Caveat: I vibe code in tiny little chunks. I see what I want to do, and exactly how I want it done, then prompt that, refine, what was output, then repeat. I bet Fable is better at building a whole app from a 2-sentence prompt; but that's just not important to me at all.

      • akmarinov 2 hours ago
        Same here - gave 5.5 a web design to implement and it sucked. Gave the same to Fable and it still sucked.
        • AgentMasterRace 1 hour ago
          did you use Claude design, their tool meant for Web design? because if not then you're the problem .
          • akmarinov 35 minutes ago
            I did. It was still Claude that’s the problem
          • bbg2401 51 minutes ago
            Perhaps you could explain why they are the problem.
    • aroman 2 hours ago
      Claude Code fan here... Codex is very good. Sometimes better. The killer feature is price.

      After 6+ months of exclusive Claude Code usage, I was begrudgingly forced to try Codex once Anthropic rejiggered their limits such that I kept maxing out my $200/mo plan in just a few days. These days I pay both $200/mo plans, and it's just about enough to get me through a week's work (small game studio - infinite code to write!)

      • ryan_n 21 minutes ago
        Genuine question/not a critique-are you actually reviewing all that code or just sending it and hoping for the best? I just can't imagine someone is reading/reviewing that much code every day, but maybe I'm wrong?
        • aroman 13 minutes ago
          Like before AI, the scrutiny varies with the sensitivity of the area being edited.

          Simple UI change? I do an AI review, but otherwise neither read nor write the code. The models are good enough they write better UI code than me, 9 out of 10 times. Not always the more idiomatic, but usually safer and more correct.

          Change to our core data plane? I might spend 2-3 times more effort reviewing it than before AI. Yes, I go more slowly than pre-AI. Many more reviews, many more angles considered, including both human and (lots of) AI review cycles.

          Most code is not that critical, and AI is also scarily good at writing tests. We also spend considerably more time paying down tech debt and testing thanks to AI, now that the cost is near-zero.

          Net: I spend 10-25X less time on low-risk changes. I often direct (or at least approve) the implementation approach, but I rarely read this code. I spend 2-3X more time on high-risk changes. In both cases, I never write code "by hand". Since about November, I've had no reason to actually edit code in a code editor (perhaps maybe except .env files, which we don't allow agents to edit for obvious reasons).

          AI is a tool. You can use it to go fast recklessly, or you can use it to go slow with confidence. Just like before AI... the skill and art of engineering is knowing when to do which.

      • ValentineC 58 minutes ago
        > (small game studio - infinite code to write!)

        Curious: what multiplier do you think your productivity has increased by, from before AI?

        • aroman 20 minutes ago
          In terms of ability to ship? Easily tenfold. We literally ship 10 times more than before AI. This does not, however, translate into a tenfold increase in actual business success, of course :)
    • steve-atx-7600 38 minutes ago
      Set yourself up to be able to try / switch between models easily. I was a claude only user and just have my user level AGENTS.md for codex and others simply point at my user CLAUDE.md. Have a script that syncs my skills (just directories) between all models. Also, if you want to use /simplify or similar from claude in another model, you can ask claude for the prompt and put that in a skill for the other models.
    • InsideOutSanta 2 hours ago
      > does it really matter anymore?

      They're different models with different philosophies behind them. This is anecdotal with a user group of 1, but in my experience:

      Claude has a stronger personality and is more creative. If you give it vague instructions, it's better at filling in the blanks with reasonable ideas.

      GPT-5.5 is better at following instructions. If you know exactly what you want, it will do it without going off the rails. It's also less likely to imply that you're dumb, but I don't really care about that. Some people do.

      • akmarinov 2 hours ago
        I’ve found that Claude is very literal. When I talk to 5.5 it gets what i want it to do, when I talk to Opus 4.8 it does what I say literally and doesn’t get the intent behind it.
    • kristianc 3 hours ago
      Codex has been good for a long time, more expensive but very focused on efficiency. Working with it feels faster and more to the point than Opus models and I trust it more with long-running jobs. Also regular resets vs being at the whim of Anthropic drama all the time is hella nice.
      • anukin 3 hours ago
        Codex is cheaper on average no? I think the models are expensive but the token efficiency of the harness itself solves the problem.
        • kristianc 2 hours ago
          Yes that's what I meant, the per token cost is higher but as you say the efficiency levels it out / works slightly in Codex's favour.
      • timcobb 2 hours ago
        Anyone know what the deal is with the resets?
        • kristianc 2 hours ago
          They've discovered it's a good marketing strategy. Whenever there's an outage, or a new launch, there's often a reset with it, which helps keep people engaged with OAI / Tibo and reduces churn.

          They've also introduced banked resets, which are really clever. If you have a $200/month plan and three banked resets, you're not churning because you will overweight giving up those resets (loss aversion theory).

          • timcobb 2 hours ago
            I ran out of resets :( hehe I had 3 and used them all
    • CuriouslyC 3 hours ago
      It never really mattered (except when codex was very new). If anything, codex's remote session integration is better, so outside of some "ultracode" orchestration bells/whistles where Claude Code is ahead, I think Codex is a better tool.
      • mountainriver 2 hours ago
        Agree, I think there was just a blind study that showed no one could tell the difference even though the users were avid they could
    • alexhans 2 hours ago
      > What's the consensus today on codex vs claude code, does it really matter anymore?

      Consensus is probably the wrong word for the popular opinions reflected in HN that you might get.

      I would recommend that you have 2 of each at all times when it comes to AI so you don't necessarily become overly locked to quirks of one thing. You'll soon realize that things move so fast that you just start internalizing common patterns instead of depending on one specific vendor.

      I recommend that you try pi and codex besides claude, to get your own feel for it.

    • Daedren 3 hours ago
      Use a harness that doesn't lock you into a moat, like OpenCode.
      • magicalhippo 1 hour ago
        You can use Codex with any endpoint compatible with OpenAI Response API[1], like llama.cpp.

        [1]: https://unsloth.ai/docs/basics/codex

      • thebigspacefuck 3 hours ago
        FWIW Claude Code works with OpenRouter so you can use any model.
        • theturtletalks 2 hours ago
          CC system prompt is bloated, use Pi to test Codex instead
      • akmarinov 2 hours ago
        Codex is open source and lets you use any model https://learn.chatgpt.com/docs/config-file/config-advanced#o...
      • maxloh 2 hours ago
        Codex CLI is open source too. I don't think there is a difference.
      • AntonyGarand 3 hours ago
        Can't use a claude code subscription in another harness though
        • I_am_tiberius 3 hours ago
          Don't use providers that don't allow it.
        • kristianc 2 hours ago
          You can however for now use wrappers which are not harnesses such as T3Code though. They were going to cut under the Programmatic API, but have at least temporarily walked it back.
        • greenavocado 3 hours ago
          You absolutely can; they are not banning anymore. The bigger problem is that subscription versions of the models are way crappier than when the "same" model is hit via API (Bedrock/Vertex)

          You can also make it not count against extra usage.

          OpenCode docs show it because Anthropic specifically ambushed them with a PR to remove support so simpletons can't use it easily.

          • AntonyGarand 3 hours ago
            The opencode docs[0] still say otherwise, do you have a source?

            [0] https://opencode.ai/docs/providers/#anthropic

          • ghostpepper 2 hours ago
            I am curious about the claim that the subscription models are different. Has anyone benchmarked this?
            • retinaros 1 hour ago
              maybe he meant about server-side features? otherwise its dumb. OAI and Anthropic are using azure/aws/gcp...
              • AlexCoventry 22 minutes ago
                How does that work? Doesn't that mean Microsoft/Amazon/Google have full de facto access to OAI's and Anthropic's model weights and operational processes?

                This is one reason it surprised me that Anthropic decided to run stuff on Musk's hardware. It seems overwhelmingly likely that the new Grok release is informed by what Musk has been able to learn from that relationship.

          • infberg 3 hours ago
            Do you have a source for that?
          • llm_nerd 3 hours ago
            They aren't banning it anymore, they just make it count as "extra usage". e.g. you're paying for every token in addition to your subscription.

            Further, the claim that the subscription "version" of the model is worse sounds like bullshit (and the sort of anecdotal nonsense that you see on sites like this). Do you have anything substantiating this?

    • xur17 2 hours ago
      I personally use opencode so I can swap between models and try different options. I'd say I prefer claude (fable and opus 4.8) so far, but curious to see where gpt 5.6 lands.

      For personal stuff, I've been pretty happy with chatgpt's $20 plan. I believe it has considerably higher limits than claude's $20 plan, and it's enough for the personal stuff I play with (hermes, and some small coding stuff). Also allows me to keep up to date on openai models.

      • timcobb 2 hours ago
        The $20 GPT plan with GPT 5.5 lasted me, somehow, exactly one smallish fixup feature
        • xur17 10 minutes ago
          Which limit? Weekly or 5 hour?

          I've been using it with hermes and some coding (with opencode), and I am getting a LOT more than one feature out of it, but the work is spread throughout the week.

    • bakies 3 hours ago
      I spent the last couple days switching because Anthropic keeps locking stuff behind API pricing. OpenAI lets you do anything with your sub right now. I'm building headless and web interfaces around Pi.dev. I had this previously with Claude Code but they are going to lock away all those features. I think the Claude does a better job at being proactive to solving things, but I'm going to keep tweaking my harness to nudge gpt to do more in it's turn. Not sure!
    • edumucelli 2 hours ago
      Not sure about the consensus, but during an entire week I have done every task on my workplace with both Opus 4.8 and GPT 5.5. GPT won hands down. I would even sometimes copy the plans and solutions (using different Git worktrees) from GPT and paste it on Opus and itself would say GPT plans were better. At that point I have migrated. Fable is not enabled in our workspace so I have not tried.

      Claude lost my trust around February this year when the plan would say nonsensical things as "delete this method" that was clearly a key method on that part of the codebase.

      For personal projects I am using Codex 20$ plan and when that is over I use DeepSeek which is insanely good for the cost.

    • sidrag22 3 hours ago
      Personally, I started using openai models to mess with other harnesses. I was pretty oppositional to CC and how they don't let you kinda plug and play freely, or give transparency into -p usage with other harnesses. So i mix and match a bunch of openai and some chinese models im trying out into opencode. I keep hearing codex is great, on the tier of current CC, I've tried it and it just ate my entire 5 hour usage window looping without asking for clarification on something and none of it was usable. that was the only time i tried codex as i could got that same task done with maybe 20% of my window with my existing openai opencode workflow.

      I had put a decent amount of effort into setting up that initial codex attempt and it went so poorly that i've been entirely uninterested in trying again. This was maybe a month or so ago, and i know stuff moves fast, but for me, i like the models, dont care for the harness.

    • teki_one 1 hour ago
      I had great results combining the two. If you (or your employer) can afford then you can ping-pong the models in the plan phase (not really ping-pong as humans should get a say too) and then let one implement and the other review. I got better results working this way than just to stick to a single model.
    • vessenes 3 hours ago
      More literal, less fluid verbally, harder time understanding nuance, more correct code, fewer bugs. Less pretty UI. I switch back and forth but find I have less 'clean up' work with codex; more upfront communication though to properly specify. High hopes for 5.6!
    • Skidaddle 43 minutes ago
      A few less obvious niceties of Codex:

      - built-in image generation using your subscription, which can be super handy

      - can actually edit Google Docs and Google Sheets (Claude can only create new or sometimes append)

      - I get a surprising amount of mileage out of the $20 plan

      They both have their places for sure.

    • NiekvdMaas 3 hours ago
      In my experience, for coding Codex is definitely far ahead of Claude Code, even when using Fable 5 as a model.
      • postflopclarity 2 hours ago
        you have a very strange experience
        • arcanemachiner 1 hour ago
          People work on different things, fail to mention their field of usage in the comments, and then misunderstand the experiences of others who do the same. Repeat ad nauseam.
        • AgentMasterRace 1 hour ago
          he's writing novels
    • lizardking 1 hour ago
      I consistently have better results with Codex for the work that I do. People have been saying that for six months, but until 5.4 the experience was sufficiently slower that it wasn't worth the switch. Making the switch was frictionless. Give it a try
    • osigurdson 2 hours ago
      I use both. Not because I am cool, but because it is cost effective for personal projects with two $20 / month plans. It is also nice to be able to see what the state of the art is like for both.

      Personally, I find it very interchangeable. I open codex --yolo or claude with whatever there yolo flag is (have an alias).

    • yokoprime 1 hour ago
      I prefer codex for most tasks, but stil use Claude if i need to make something "nice but generic", i.e. a html artefact or touch up of front end code.
    • corford 1 hour ago
      Personally I use Open Code with a copilot sub. Then all models are available in my session with just a /model and /variants command combo. Makes it super low friction to try different models & combos (my favourite right now is DeepSeek V4 Flash for initial PRD then Fable 5 high for implementation).
    • linsomniac 1 hour ago
      I'm also a long-time Claude Code user here, though the last 3 weeks I've been doing loops having claude use codex to review until they reach consensus; uses tons of tokens but the result is really good.

      I'm trying Codex as my primary the last day or so, because I'm at 98% use and reset in 3 days on Claude. I'm worried about a lot of our skills and CLAUDE.mds and the like getting lost unless I migrate them, but otherwise codex seems to be working great.

    • nvarsj 1 hour ago
      The harness is so much better than cc which is a buggy mess. Gpt is also way faster than Claude. I’ve been using gpt for a while now and I know a lot of people that swapped away from Anthropic for multiple reasons. However - fable still seems to be the best coding agent, it’s just slow and the harness sucks. So I still use it in some rare cases like to review codex. I’m hoping 5.6 lets me drop it entirely.
    • dgritsko 3 hours ago
      I use Conductor pretty much exclusively and it makes it incredibly easy to try different models, even within the same workspace - definitely recommend giving it a shot. Whenever I'm forced to use the Claude Code app directly it just seems woefully inadequate compared to Conductor
    • ThunderBee 1 hour ago
      IME it entirely depends on your work. I find myself using both daily for different things.

      Codex with GPT 5.5 is much better at general SWE tasks but Claude Code with Opus is far better at complex reasoning tasks like reading and summarizing research papers, replicating experiments, identifying research gaps and proposing interesting follow ups.

    • timcobb 3 hours ago
      If you can afford it and you have something to justify the expense, I would get both. they're interesting to run side by side, you can hand things off from one to the other. Pretty neat. Unfortunately now I just want to have both :(
    • 3371 2 hours ago
      My experience is that Codex's auto review is extremely costly, with $20 on both sides, I can run CC with auto mode for longer than with Codex's auto review enabled. Also in my own experience Claude's usage is actually bigger than Codex, but I am not sure if that's due to I stick to 5.5 with Codex while keep Sonnet as the default to orchestrate other models in CC.
    • jghn 2 hours ago
      I have found Claude Code to be so much better than other common harnesses that it's kept me solely in the Anthropic ecosystem.
    • novaleaf 2 hours ago
      I sub both codex and claude at 20x. I like opus+fable more than gpt5.5 because it seems gpt tries to finish tasks by leaving any ambiguity unresolved. claude seems better at surfacing open questions.

      This is using the same AGENTS.md prompts, which were designed firstly for Claude use, so maybe it's something that could be optimized better if I understood gpt as well?

    • sk4rekr0w 3 hours ago
      Codex app is a much different experience than CC CLI. I would try it out for a couple days with the new model suite and see what you prefer after that.
    • urams 2 hours ago
      If you can afford to test it seriously, running both in parallel, it's worth a test to see which you prefer. If you can't, don't bother. You're not likely missing anything since they are close to personal preference with most people I know who have meaningfully tried both preferring Claude
    • athrowaway3z 2 hours ago
      They blocked Claude from being used in a different harness as well squeezed the usage like crazy. Switched to Codex and haven't cared since.

      Between the two the biggest difference by far is ... getting your harness / AGENTS.md / skills / tools set up right.

    • siva7 1 hour ago
      Have been long time clauder but honestly codex feels much more liberating. Something you can't buy..
    • giancarlostoro 2 hours ago
      Last time I tested Codex on a cheap plan, it barely lasted an hour? I think this was for the $20 plan. I was afraid to try the more expensive plan after that. Not sure, I might just outright rip my Claude Code bandaid if the current usage quotas do die off after the 17th or whatever date they said they would "return on".
    • indy 3 hours ago
      You wouldn't be leaving Claude Code, just trying something new. If you don't like it just resume using Claude.
    • purpleidea 1 hour ago
      The codex software is garbage compared to Claude, but open source is the future, so you should at least switch.
    • YuechenLi 2 hours ago
      The answer is it depends. Claude's generally better at frontend and debugging tasks, while Codex is stronger at backend features and exploratory work. They have very different coding styles and thus very different strengths.
      • TranquilMarmot 2 hours ago
        Any actual data backing this up? Or is this just your personal experience?
    • cmrdporcupine 3 hours ago
      I left Claude for Codex months ago. I was an early Claude Code adopter but I have found Codex consistently better since about the February time frame. And far more reliable.

      It's more diligent and empirical and results focused, and less creative. It sometimes needs a kick to avoid a Zeno's paradox of incremental steps to get to the goal. But it produces more reliable code with fewer race conditions, unhandled negative cases, etc.

      It's also better value from a $$ POV, or at least has been. This fluctuates a bit.

      You're also free to use your Codex subscription with other harnesses, like opencode, etc. Unlike Anthropic. Plays better with others.

    • rib3ye 2 hours ago
      It's not clear replies to this thread aren't openAI employees or incentivized influencers, but every benchmark has gpt-5.5 underperforming opus 4.8, sometimes by as much as 10%.

      Can they all be wrong/paid-off?

    • catketch 3 hours ago
      Not sure there's going to be a consensus, but I can tell you that when i have codex review claude-written code, it finds important gaps and fixes. The reverse is also true. Both are powerful, but even better when used in combination
    • thebigspacefuck 3 hours ago
      IMO LMArena is the best benchmark that avoids benchmaxxing

      https://arena.ai/leaderboard/agent

      5.6 isn’t on there yet but Fable leads by a significant margin atm

      • TranquilMarmot 2 hours ago
        The results here match up to my real-world experience using these models every day at work and switching between them regularly.
    • setnone 1 hour ago
      like others said in the thread: much less drama and i'll add much less attitude from the company and the models, overall i'm having much calmer experience with codex, hope it stays that way
    • arikrahman 3 hours ago
      For me the biggest shift was using Deepseek through an American provider with reasonix as the harness, making cache hits at a rate of practically free.
    • mountainriver 2 hours ago
      There was just a study showing that when presented blindly no one could tell the difference yet users were avid they could
      • hk__2 2 hours ago
        There _is_ a difference in the way Claude and GPT write. Last Friday I felt Opus was becoming dumb because it was writing like GPT.
    • maxloh 2 hours ago
      I wish they open source their desktop app and built-in skills one day. That would be a final blow for me.
    • qaq 2 hours ago
      I use both especially for checking each others work. Pretty happy with results
    • saberience 3 hours ago
      They are both excellent but excel in different areas. Fable is super super proactive and great for doing a LOT of work with a single prompt, also for creative work.

      Codex is more details focused, often catches wonky bugs and correctness issues that Fable misses, feels more terse and less "friendly", more like a stern senior engineer versus a friendly talkative engineer (Claude). Codex is also better if you're already an engineer, Claude is better for non-engineers. I.e. Codex works better if you know exactly what you want and know the right way of explaining it.

    • firemelt 1 hour ago
      just try it you will back to codex because gpt is trash, I ask for refund under 7 hours
    • 217 3 hours ago
      Consensus itself does NOT matter, omp is objectively the best harness for power users yet it has 0 hn posts about it, zero.

      You're fully free to use and try anything and without caring about what others think is right

      • trollbridge 2 hours ago
        omp is really good.

        I have one non technical people in my firm using it. One is using it to assist with editing books, basically using it to gather up manuscripts from e-mail / Google Doc etc. submissions, and then switch models between a cheap one and Opus (for actually analysing the manuscript).

        The other non-technical person has done really surprising things with it AI, like a long-running GPT 5.5 Pro chat session which is basically her expense tracker - it has an .xlsx file "carried" in the chat, and she just tells ChatGPT (or scans a receipt) whenever she has a new expense, and then prompts it in natural language when she needs a report. I'm looking forward to seeing what she can do with omp.

      • ljm 2 hours ago
        I figured pi itself would be the best harness because it's barebones and you make it what you want. omp is to pi what doom is to emacs is what lazyvim is to neovim.
      • Avicebron 2 hours ago
        The fact that I thought that this was amp misspelled until i someone validate omp and the checked myself indicates it's a subjective assertion at best.
      • saberience 2 hours ago
        How can this be "objective"? Surely its subjective.

        I've tried a fuck load of harnesses but keep coming back to Codex as my harness.

      • financltravsty 2 hours ago
        Absolutely. It's the only harness that is actually RSI and not run by idiots.
      • joe_mamba 2 hours ago
        > omp is objectively the best harness for power users

        Care to detail this?

      • efficax 2 hours ago
        "objectively the best"?
        • incognito124 36 minutes ago
          To quote a friend:

          > Well it's objective _to me_

        • saberience 2 hours ago
          Probably means subjectively according to his own opinion...
    • prospector1065 3 hours ago
      Try OpenCode and you can point it at either model
    • wahnfrieden 3 hours ago
      With the exception of Fable which is going away anyway, Codex is better especially after the last couple Opus releases. It’s also no longer slower than Claude.

      You get much more generous usage from the 20x plan.

      And you get far better uptime.

      If benchmarks and early tester impressions are accurate, you also get access to Fable level capability at greater speed and lower cost (included in subscription).

      • petesergeant 2 hours ago
        > Fable which is going away anyway

        $2 says nah. You can't take Fable away in a week where GPT-5.6 and Grok 4.5 launch, if you want to hold on to customers.

        • mortenjorck 2 hours ago
          The fact that they already extended subscription Fable once would suggest it won’t be solely locked behind API next week, but at the same time it really does look like they are doing everything they can to avoid serving it continuously at scale.

          Knowing Anthropic, this unfortunately might end up meaning a quietly quantized Fable on subscription.

          • wild_egg 1 hour ago
            Can anyone explain this "quietly quantized" model idea to me from a business perspective?

            Coca-Cola doesn't "quietly water down" its product to save a few bucks. They know people will take a sip, say "oh that's not what i wanted", and go buy a Pepsi.

            If they serve me a quantized Fable, I'm just going to think Fable sucks and go get my tokens elsewhere. What's the point?

            • AlexCoventry 9 minutes ago
              How are you going to confidently tell the difference between it being slightly dumber, vs just having bad luck with your recent attempts?
            • ValentineC 53 minutes ago
              > Coca-Cola doesn't "quietly water down" its product to save a few bucks. They know people will take a sip, say "oh that's not what i wanted", and go buy a Pepsi.

              Coca-Cola is also mostly measurable and reverse-engineerable.

              The Claude models are black boxes, and actively curtail distilling efforts.

            • petesergeant 1 hour ago
              Pepsi may not water down its product, but your local diner may well decide to put a little less ice-cream in its milkshakes if it thinks it can get away with it and most people won't be able to tell.
    • erichocean 2 hours ago
      I use Claude for planning, writing CRs, and code review.

      Codex writes all of the code, no exceptions.

      Works great, especially when you ask Claude to break up large CRs into roughly 10 minutes of Codex work each.

      • silksowed 1 hour ago
        Same here. I find the design, architecture, system design discussion to be better on Claude, but after Opus 4.6 I switched over to Codex for actual coding and love the results. I use both via the CLI and generally tell Claude to output the result of our decisions as a markdown that will be easy to read and implement by an agentic coding tool. Then I fire up Codex and read said markdown as the input of the session and way to build all the appropriate context needed. I see this as a way to step into letting the agents go run on their own and interact with each other, but I still like to steer so I put these manual steps in the flow. Letting the agents go off on their own and one shot big chunks is not reliable enough yet imo.
    • znpy 2 hours ago
      > I'm hesitant to leave Claude Code behind for something new.

      Codex and Claude Code are not mutually exclusive, you can use both.

    • petesergeant 2 hours ago
      In my projects, Claude writes and Codex reviews, and I've had a lot of code I've been very happy with out of that, although as of today, Grok _also_ reviews, and finds interesting new stuff.
    • simianwords 3 hours ago
      Codex UI is way way way better than Claude Code

      - codex UI is much more responsive

      - i get feedback about the progress easily

      - the tool calls and results are very legible, I can click them and see the progress

      - no one talks about this but the tool call and response notification are handled much more elegantly in Codex. In Claude Code, it is handled in a clunky way using loops which always causes some delay

      - you can steer the conversation midway in Codex

      - /side is underrated (/btw is the equivalent and is much worse in Claude Code)

      - I have to admit subagents are handled better in Claude Code

    • ralusek 3 hours ago
      I use both constantly for different things. You don't need to be a one-model Andy
    • wyre 3 hours ago
      Claude Code is a massively bloated agent harness.

      Try Pi: https://pi.dev/

      • blovescoffee 2 hours ago
        Pi is so “unbloated” that it’s extra effort to use. You can decide how much work to put into it. I get the trade off. But this is a big jump from CC. I’d recommend some middle ground like opencode.
        • lubujackson 2 hours ago
          Even simpler, use Cursor with any frontier model. I see others sweat to add enough context to Claude Code while Cursor has a ton of contextual awareness, uses subagents automatically and is significantly faster with no drop off I have found. I'm not sure why devs are so enamored with living in the CLI, but Cursor has one of those too.
          • blovescoffee 1 hour ago
            Similar to running arch Linux. Many people do need to. But many people just like tinkering. Tinkering can lead to positive outcomes but it’s usually not “doing work”.
        • trollbridge 2 hours ago
          It's worth trying out OpenCode, then oh-my-pi, and also the commercial harnesses like Codex. (I haven't yet bothered to try Antigravity and have no interest in Gemini-cli now that it's not available except on expensive plans.)

          pi is also worth tinkering with, particularly if you have an eye towards automating some things.

      • AgentMasterRace 1 hour ago
        better, use oh my pi.
    • marknutter 1 hour ago
      It honestly baffles me how people can ask a question like this and get such a wide spectrum of answers in response. It's all so much based on vibes and anecdotal evidence. I've not really noticed much of a difference in capability since Opus 4.6 and I've used a ton of different models. They all work pretty damn well for me.
    • Razengan 2 hours ago
      I've subscribed to ChatGPT/Codex for over a year and tried a Claude sub twice 1 month each, with a gap of several months in between.

      I tried them both side by side, mostly for reviewing existing Godot/GDScript code, or sometimes generating Swift Mac apps, including converting ancient relics I wrote eons ago in Visual Basic on Windows

      Codex was consistently better than Claude: https://i.imgur.com/jYawPDY.png

      Besides the useless "This is good" findings while reviewing and the excessive "oops you're right" backtracking, Claude's atrocious UX and borderline "spyware" make me never want to try an Anthropic product again for a long long while.

    • moomoo11 3 hours ago
      [dead]
    • Jtarii 3 hours ago
      Literally every top model is identical and anyone saying otherwise is engaging in astrology.
      • Supermancho 2 hours ago
        The outputs, ui, and overall behavior (tokenization) are not identical.
      • postflopclarity 2 hours ago
        anybody saying they're identical clearly doesn't use both...
        • iugtmkbdfil834 2 hours ago
          Honestly, I would even push it further. People who would claim that don't use either one.
  • eig 3 hours ago
    Funny to see that they did not include Fable 5 in their GeneBench and LifeSciBench comparisons because "it does not answer advanced biology questions and refuses the majority of questions in this eval".

    Winner by default!

    • inciampati 2 hours ago
      This is a major reason why I and a number of biologists I've talked to have canceled their anthropic accounts recently. Not working is not working.
    • midtake 1 hour ago
      You shouldn't know too much about biology, stupid human. You might live your life in an unexploitable way.
    • paxys 3 hours ago
      Where’s the lie?
  • senko 44 minutes ago
    I love testing the new models by asking them to code a toy RTS game. Here's what Terra did: https://senko.net/vibecode-bench/2026/rts-gpt-5.6-terra.html (one try, in codex app, xhigh effort)

    Comparing this to other models, I find it similar to GPT-5.5 and a bit behind Sonnet 5. You can see how other models fared here: https://senko.net/vibecode-bench/ (you can also fetch the prompt and the the 5.6 Terra resulting code on from that page).

    I don't have access to Sol yet (on a Plus sub, which should get it according to what I've read), so can't do the more interesting test. I'll update the above page as soon as I get access - hopefully soon.

    • gorgmah 6 minutes ago
      this is so cool: it's playable (even though super boring since there are no enemies) and you can feel that a few iterations would make it very usable.

      Which model is the best at the moment, for this kind of stuff, in your experience?

    • rsoto2 13 minutes ago
      So the measure of a model is how well they can recreate something they easily have thousands of examples of in their training data. There's probably a better base RTS on github somewhere for free.
  • mchinen 3 hours ago
    The frontier graph on all these benchmark are extremely in favor of 5.6 Sol over Fable, more than the best model comparisons in previous iterations.

    I'd like to know how cherry-picked this is, and what tests it performed less overwhelmingly in, but I suppose that info is not going to be on this post.

    If it pans out to be as good as it says, that's great. On the other hand, if this model is not overwhelmingly impressive over Fable, I will lose what remaining trust I had in these announcements.

    • thurn 2 hours ago
      They do disclose that they scored much lower than Fable on SWEBench Pro, which is a pretty high-quality benchmark. I think it's partially just about what they choose to emphasize...
      • danielsamuels 2 hours ago
        It's worth noting that OpenAI recently came out saying, "We don't think SWEBench Pro is worth reporting any more" - https://openai.com/index/separating-signal-from-noise-coding...
      • bel8 2 hours ago
        SWEBench Pro should be ignored until they fix it or disprove the broken task accusations.
      • futureshock 1 hour ago
        There has been a lot of chatter ever since the Mythos scores had been release that SWEbench pro had major contamination and that Mythos had memorized many questions that lacked the context to be solvable on their own. And now with OpenAI saying a large number of the questions are broken, I think it's worth taking that single outlier benchmark with some salt when the overall trend is that 5.6 is very competitive with Mythos at about half the price.
      • mchinen 2 hours ago
        I totally missed that, because in the charts they showcase for coding, the SWEBench score is not present, they only include it at the end of the post in tables. Hmm.

        Great catch.

        • saberience 2 hours ago
          The SWEBench benchmarks are really gamed at this point and should not be trusted period. The solutions are effectively in the training sets and have been for a while.
          • mnicky 1 hour ago
            SWE Bench Pro is completely different benchmark than SWE Bench (e.g. Verified) suite was. It only copied the name.
            • saberience 3 minutes ago
              SWE Bench Pro is also gamed and shouldn't be trusted.
      • simianwords 2 hours ago
        > SWEBench Pro, which is a pretty high-quality benchmark

        No, doesn't seem like it

        https://openai.com/index/separating-signal-from-noise-coding...

      • lwansbrough 2 hours ago
        Didn't they also just post about how SWEBench is broken?
    • therobots927 3 hours ago
      The proof is in the pudding and these benchmark stats will only work for so long before people lose interest.
    • enraged_camel 2 hours ago
      The charts are also extremely difficult to parse. They seem auto-generated. Dataset coloring is atrocious.

      Regarding your main point, yes, I agree. My impression (as someone who uses both Codex and Claude Code daily) is that OpenAI does a fair amount of benchmaxxing.

  • Jcampuzano2 2 hours ago
    I really wish there was just an easy guide on when to use Sol vs Terra vs Luna, and it just moves further into confusing territory when it comes to naming.

    The naming convention is especially difficult to decipher depending on what your native language is. Of course a latin language speaker might be able to easily determine oh yeah each one is slightly bigger than the other but I still think it borderlines too confusing.

    That aside all the numbers look amazing, and I'll be happy to probably main this alongside grok-4.5 for a while comparing the two on price and efficiency.

    I vastly prefer the direction that OpenAI seems to be going with token efficiency and performance compared to Anthropic who seems to be moving towards a world where you just token-max as much as possible ignoring any and all costs.

    • jstummbillig 2 hours ago
      Why would you need a guide for that now? We long had to pick different models (and thinking levels) by task and feel.
      • Jcampuzano2 2 hours ago
        Previously it was much more obvious which model to reach for depending on your use case because they had the mini and nano naming conventions.

        Getting rid of that seems like a step back. Just a personal nit though.

        I've seen buzz about this elsewhere as well but to me effort levels seem more like spend limits disguised with another word. I don't think they should even exist.

      • aerhardt 1 hour ago
        My guide was to pick the best model on "High" for 99% of tasks.
      • bigyabai 2 hours ago
        The naming convention is bizarre and doesn't really mean anything to normies. Trying to pick between "Sol" and "Terra" is like asking the average person if they want the Max or the Ultra chip.
        • minimaxir 2 hours ago
          What about Haiku, Sonnet, and Opus?
        • slekker 2 hours ago
          The sun is bigger than earth which is bigger than the moon, it's pretty simple really
          • bigyabai 2 hours ago
            Which is cheaper to use? The size euphemism is a really roundabout description vs "Nano" and "Pro" for the layperson.
    • jrflo 58 minutes ago
      My guess is that it's the same for Haiku/Sonnet/Opus: Biggest model for architecture and high level planning and technically challenging problems, medium model for simple implementation tasks, small model is for nothing
    • asadm 2 hours ago
      it's simple: unless trivial TOIL, always use the highest at ultra max settings.
      • paxys 2 hours ago
        Okay Richie Rich
    • taytus 2 hours ago
      You don’t know what sol means? You don’t understand the difference in sizes between Terra and sol? I’m genuinely asking.
      • Someone1234 2 hours ago
        That isn't what "genuinely asking" looks like, you're criticizing using "questions" as cover. It isn't subtle, nor is it constructive.

        I agree with them, Sol, Terra, and Luna are confusing names. They mean the same thing as GPT-5.6-Max, GPT-5.6-Plus, and GPT-5.6-Fast but require base knowledge for an analogy.

        It feels like it was adding by the marketing department.

        • HarHarVeryFunny 31 minutes ago
          Surely it's size based: Sun (Sol) > Earth (Terra) > Moon (Luna)

          Similar to Anthropic's size/length based naming: Opus > Sonnet > Haiku

          These names seem easy to understand to me, and much clearer than suffixes like -max and -plus.

        • LUmBULtERA 1 hour ago
          >They mean the same thing as GPT-5.6-Max, GPT-5.6-Plus, and GPT-5.6-Fast but require base knowledge for an analogy.

          But do they though? When do you use GPT-5.6-Max-Low vs. GPT-5.6-Plus High? Or GPT-5.6-Fast-Xhigh? What's the Pareto optimal choice (outcome and price)? According to the benches it seems to bop around and the even if the benches are accurate the best choice isn't always consistent.

          • Someone1234 1 hour ago
            > When do you use GPT-5.6-Max-Low vs. GPT-5.6-Plus High?

            You don't, because that isn't something I proposed using for model naming.

            I called them GPT-5.6-Max, GPT-5.6-Plus, and GPT-5.6-Fast. Reasoning levels are distinct from the model design itself, and the UI makes that clear.

            Plus, using that same flawed argument this would be called GPT-5.6-Sol-Low or GPT-5.6-Luna-High which also makes no sense/is confusing. So that argument applies (or more accurately doesn't), no matter the model names.

      • Jcampuzano2 2 hours ago
        Did you not read the second sentence? Obviously I know what sol is given my first language being Spanish. I'm just speaking in a general sense that it can be confusing for others.

        I already know plenty who had no clue what the difference between Terra and Luna would be.

        • ealready_value 2 hours ago
          My first instinct was Sol > Luna > Terra, since Sol is the farthest away, then Luna, and Terra is the closest. Size was not my first instinct. Or should Terra be the best model because its closest to people, then Luna because there have been people on it, then Sol be the worst because no human has been there?
          • endemic 2 hours ago
            The naming scheme is too "clever."
      • adamrezich 2 hours ago
        Sure—so, is Sol 109.2x better than Terra? Or 1.304x10^6 better?
  • aliasxneo 2 hours ago
    "We've extended usage of Claude Fable" message incoming any day now.
    • ftchd 2 hours ago
      They reset all usage half an hour ago. It's back to 0% per week and session. No specifically Fable related.
      • halfmatthalfcat 1 hour ago
        Hahaha seeing this play out in real time is absolutely incredible.
        • danielbln 1 hour ago
          Im here for it, good on Anthropomorphic to feel some heat again after all that drug dealer Fable business.
          • aliasxneo 57 minutes ago
            100. I'm so tired of being treated like a drug addict by them. I'm currently sitting on _four_ "reset vouchers" from OpenAI so I get basically all week to play with 5.6 Sol to my hearts content. The amount of positive sentiment that brings to me should really be a concern for Anthropic who is increasingly alienating me away with their shit strategies.
  • tekacs 9 minutes ago
    Unfortunately, I'm finding that in long-form agentic use, when I'm trying to use Sol, I keep tripping guardrails – moreso than even Fable, somehow.

    I don't know exactly what part of my codebase is triggering it, so I'm going to have to keep poking, but apparently the guardrails are not that gentle despite the phrasing. :(

    • ramijames 2 minutes ago
      Sounds like you are working on something naughty :)
  • mNovak 2 hours ago
    >> approximately 700,000 A100e GPU hours of black-box automated red teaming

    Amusing that they use A100e as the reference point to sound impressive. Different ways you could make that conversion, but based on FP4 FLOPs (yes it's disadvantageous to A100, that's the point), that's something like 200hr on a GB300 NVL72 rack.

    Not nothing either, but far less astounding sounding than 700k hrs.

    • BoorishBears 1 hour ago
      I'm pretty sure Altman has spoken about giving a model 100k+ A100s specifically, this might be them being very literal
    • stavros 2 hours ago
      Wait, what do you mean? 700k A100e hours are equal to 200 hours of a GB300 NVL72 rack? One GB300 NVL72, 72-GPU rack has equal processing power to 3500 A100e GPUs?
      • pizzafeelsright 1 hour ago
        maybe? ai says about *8.3 days* of continuous runtime on a single GB300 NVL72 rack

        about a sprint's level of effort.

      • BoorishBears 1 hour ago
        > based on FP4 FLOPs (yes it's disadvantageous to A100, that's the point)

        The A100 doesn't have hardware FP4, and you'd be running a quantized model with some accuracy loss but unless this was natively trained on FP4*

        * to add another layer, they own the model and could apply tons of post-training techniques to reduce that accuracy loss and probably already do

  • arizen 3 hours ago
    "GPT‑5.6 delivers a step change in design judgment. With only high-level direction, GPT‑5.6 creates tasteful, ergonomic, and functional interfaces. Its stronger computer-use capabilities let it inspect and refine the rendered result—not just generate the underlying code or content—so it can catch visual and functional issues and apply finishing touches before handing the work back."

    This one is really promising, as it may allow to close major gap with Claude in design/UI skills

    • silksowed 1 hour ago
      Computer-use is a big limitation that my 2015 Macbook Pro cannot handle. I find the Codex cli says it looks at the end output artifact but so often it fails to refine it into acceptable form. If it could use my computer screen and visual inputs for review, it might be able to actually design documents/powerpoints/etc. I'm juicing everything I can out of the 11 year old laptop and I'm honestly impressed at what it can still do.
    • HyperL0gi 1 hour ago
      +1. I've been only using Sonnet/Opus these days for UI work because GPT 5.5 just can't do any of that. Its just really terrible. Eager to give this one a try.
    • GenerWork 2 hours ago
      Agreed, I’m looking forward to trying it out. I think that the rise of visual design skills that are pretty clearly targeted towards Codex users has lit a bit of a fire under their butts.
  • goodmattg 1 hour ago
    I flip back and forth between whoever currently has the more powerful frontier model that isn't cost prohibitive - subscriptions only, API pricing a non-starter. Today that's Fable 5 which has been excellent, as soon as it's Sol I'll switch to that. The OAI/Anthropic harness behavior has mostly stabilized for me with consistent AGENTS.md that I sync with CLAUDE.md - I like pi (pi.dev) and have tried to build it up to get performance comparable to the two "first-party" harnesses, I'm just not there yet.

    One major sticking criteria for not going with OpenCode / pi for all of my coding is I want access to the tier-1 frontier model of the day without API pricing - e.g. afaik I can't use Fable 5 via pi harness even though I have a subscription, so for this week I'm on Claude Code. It's not the need to Fable 5 for everything, but even if I just want the marginal intelligence benefit to stress test an architecture decision, it's a safety blanket to know there isn't a ~smarter~ model I could have used. And for my use cases, the doggedness and capability of these frontier models has been insanely effective.

    My feeling is we're still in the Uber era subsidy period - the moment the subscriptions either try to lock me in longer than a month or stop OAI/Anthropic stop delivering frontier models in the subscriptions, I'm out - switching fully over to pi.dev or another OS harness and routing my token spend via OpenRouter or offloading to Qwen locally. Then I'll have to put an accurate dollar amount on frontier intelligence.

    • 2001zhaozhao 1 hour ago
      I'm working on a multi-harness IDE that supports custom agent workflows and skills that are shared between any harnesses it wraps over. I think it might prove handy for a workflow like yours.
      • goodmattg 1 hour ago
        Would it currently support Fable 5 via the restrictions Anthropic is placing on usage... because that's my major blocker
  • revolvingthrow 1 hour ago
    Benchmarks look really promising. Suspiciously good, even. I guess we’ll see soon enough.

    My question to previewers: how are the guardrails for random joe that wasn’t personally blessed by the ai pope to access the non-nerfed model? Fable is a nightmare in this regard, but I’m not sure whether 5.6 also gets a critical side-eye from the gubmint when you ask it to fix bugs in your code (you filthy hacker, you).

    • cmrdporcupine 1 hour ago
      I almost immediately ran into "This request requires additional safety checks, which can take extra time. Hang tight or retry with a faster model for a quicker response, though it may be less capable of handling complex requests."

      Which is something I've never seen with codex before, and I wasn't doing anything funky. Just writing CUDA kernels and benchmarks for them.

  • cbg0 3 hours ago
    5.6 Terra (mid tier model) as good as Fable on DeepSWE while cheaper than Opus API pricing. Seems like a homerun.
    • osti 3 hours ago
      GPT usually performs better on DeepSWE while Claude does better on FrontierCode. These two coding benchmarks are pretty much the only ones right now that's still worth taking a look at imo.
    • DetroitThrow 2 hours ago
      DeepSWE seems to strongly, strongly prefer ChatGPT models. There were also major flaws in its methodology pointed out recently, that overlap strongly with the flaws OpenAI pointed out in its SWE Verified report.

      I use both ChatGPT and Claude for engineering work on a daily basis, touching performance critical code to application backends to frontend work, and I've found that DeepSWE scores don't reflect my reality when I assess high quality output from the models/harnesses.

      Not that Opus always beats GPT 5.5., but that 5.5 is ahead of Opus on a general benchmark smells off to me.

  • lukebuehler 1 hour ago
    Very interesting: I wonder if the RL approach is diverging between Anthropic and OAI?

    I noticed that Fable uses shell tools almost exclusively (even to search and edit files), compared to previous Anthropic models.

    Having run some experiments with 5.6, I notice that it uses built-in file systems and provider native tools much more (not shell tools), compared to previous OAI models.

  • twothreeone 2 hours ago
    Wow the video is much better.. the PR spend clearly went up a lot. Mainly just showing "real people" doing "real stuff".
  • HarHarVeryFunny 26 minutes ago
    Not specific to OpenAI / Codex, but I'm curious what people are doing to protect themselves from any destructive actions by their coding agents? Just install and pray? Explicity approve all actions? Reconfigure for safety? Run in a sandbox (Docker) ?
    • user43928 2 minutes ago
      I use the auto-reviewer for actions outside the builtin sandbox.

      So far this has been rock solid, and tens of millions of developers use this setup without issue.

      It is not going to wipe our hard disks. At least I hope so. Fable and GPT 5.6 have been ever more proactive, and GPT 5.6 is automating the AppStore on my machine to download an Xcode update while I am typing this.

    • kennykartman 10 minutes ago
      Typically I just want to isolate the agent disallowing it from accessing other parts of the filesystem. Using a different user might be enough, but I typically use [bubblewrap](https://github.com/containers/bubblewrap).
    • ryan_n 25 minutes ago
      I still just explicitly approve all actions and review all code (unless it's a personal/throwaway project no one else will ever touch/use/see). I know a lot of people that run in a sandbox though. That said, I'm sure there are lots of people that just yolo it and hope for the best.
  • gorgmah 1 hour ago
    Just used terra ultra for exactly one prompt in codex and it ate through my full 5h window in about 10mns (20$ plan). The results look pretty good though. Luckily I have had my chatGPT subscription for a while and have a bunch of resets available (nice compared to anthropic).

    Assuming I take the 5x plan it would give me about an hour of active sessions with terra ultra (maybe ultra is not good value regarding tokens?), not even using Sol yet. Does everyone using codex use the 200$ plan?

    I normally use the 100$ anthropic plan and barely ever reach the usage limit.

    • jstummbillig 1 hour ago
      > maybe ultra is not good value regarding tokens?

      Well, yes, as explicitly stated on https://openai.com/index/gpt-5-6/: "ultra goes further by coordinating four agents in parallel by default, trading higher token use for stronger results and faster time-to-result on demanding tasks."

    • ssl-3 1 hour ago
      I use the $20 plan, but I don't code all day every day.

      With Codex, it is my experience that I can churn through a 5h window in no time with newer models -- especially when they're new. So I tend to use fancier models for planning, and the less-fancy models for writing code based on that plan. I switch to the fanciest model if any part of this gets stuck.

      If I've got a something big-ish to work on, I pay attention to the reset timers so I can get more of it done in one chunk.

      Models seem to slowly get better/relatively less-expensive as they age. (It isn't clear to me if that's because the cost actually goes down, or if the allotment goes up, or if things get more efficient in unseen ways, or what. OpenAI is vague AF about what we get for the $20 that we pay.)

    • altcognito 1 hour ago
      Do you know if you used sol/terra/luna?
  • sd9 3 hours ago
    I haven't tried an OpenAI model for a long time, but with Fable going to API pricing soon this might be enough to get me to try codex.
    • danielbln 1 hour ago
      Seeing how Anthropomorphic just reset usage quotas back to 0 and the other day extended Fable sub inclusion by a few days, I have a feeling they might not drop Fable out of sub after all, because like you I would most definitely take a long good look at codex at that point.
      • sd9 1 hour ago
        [dead]
  • big_toast 2 hours ago
    In the introduction video they say 5.6 Sol autonomously post-trained 5.6 Luna. Curious what this means.
    • jrflo 53 minutes ago
      /goal tune 5.6 Luna parameters until performance is maximized across all benchmarks
    • vibcdingenjoyer 2 hours ago
      Sounds like they gave it a goal to hit certain benchmarks and just let it have its way with the base Luna model.
    • 2001zhaozhao 1 hour ago
      It means OpenAI and Anthropic are now in a RSI race with each other
  • hereme888 2 hours ago
    I use 5.5 a ton. It's immediately apparent that 5.6 is truly a better model. Hope they don't lobotomize it later.
  • thimabi 2 hours ago
    I’m interested in knowing how each of GPT 5.6’s variants fare in non-English writing/translation tasks.

    GPT 5.5 has a tendency to write English calques and non-idiomatic prose in other languages. Although that can be somewhat tamed with detailed instructions and a corpus of confusing terms, the model’s output often reads like a literal translation rather than native prose. Since I notice these issues most clearly in languages I know well, it makes me reluctant to trust the model’s output in languages in which I’m less proficient.

    Ironically, ChatGPT began as a simple text-generation tool, but much of its offerings and benchmarks now focus on coding and agentic workflows, while leaving behind what made it notable in the first place.

  • macleginn 56 minutes ago
    For context, I have access to MS Copilot through my workplace. To see what it looks like, I have tried to login through https://copilot.microsoft.com/ , where I was informed that my account, although recognised, is not yet supported. However, I can get more or less the same chat window, with access to all the data, through https://m365.cloud.microsoft/ A redirect could have been useful.
  • 2001zhaozhao 1 hour ago
    Huh, a good alternative just as anthropic's 50% weekly subscription subsidy is ending this weekend. Time to see if it's benchmaxxed or actually a strong leap over GPT5.5.

    They also seem to really not care about alignment, or care about it in the wrong way. It's entirely missing in the blogpost and there are some concerning bits in the model card, seemingly treating CoT controllability as something to be "investigated" rather than the warning sign it's supposed to be.

  • bob1029 1 hour ago
    I am seeing some bugginess in testing:

      Parameter: reasoning_effort
      
      Function tools with reasoning_effort are not supported for gpt-5.6-sol in /v1/chat/completions. 
      To use function tools, use /v1/responses or set reasoning_effort to 'none'.'
    
    Official OAI .NET library. Even when I override the currently experimental [?] flag to 'none', it will still occasionally throw this error (about 5% of the time).

    I hope we aren't trying to push customers off the chat completion endpoint... Responses endpoint looks great on paper, but the business wants more visibility and control over the reasoning process than this product currently offers.

    Edit: This is broken in my VS copilot setup too.

  • stillpointlab 1 hour ago
    I can't try it since it hasn't appeared in my Codex yet, but this is is necessary from OpenAI in my opinion. Fable is just so much better at understanding broad context. I only use GPT 5.5 for straight forward easy to describe tasks, and it does crush those. But I spend a lot more time steering Codex towards good design on broad concept type tasks, ones that Fable shows sometimes surprising clarity.

    I look forward to seeing how it compares once I have access. Not getting tripped by spurious safe guard flags could be an advantage.

  • Tenoke 3 hours ago
    Is any of those comparisons about Pro vs non-Pro (Pro is only available in $100+ plans)? I am curious about that but I think Sol, Terra, Luna are different sizes of it without the Pro part, and I want to know how much worse do I have it on the $20 plan compared to if I upgrade.
  • vamsiraju 56 minutes ago
    I think 5.6 Sol is only as good as 5.5 or Opus 4.8 in terms of getting its given work done. It just has an uncanny ability to pickup more work that it can tackle next that the older models lack, or have not been trained to do before. Where folks are seeing a difference between working with Fable or 5.6 I think also boils down to this phase shift.
  • tmaly 1 hour ago
    One of my best use cases for the short duration I have fable is to use it to create the plan and acceptance test files then use GPT 5.5 Pro to do an adversarial review on the plan then feed that feedback into fable to fix the plan.
  • xur17 1 hour ago
    Looks like I have access to gpt-5.6-terra and luna. How does one decide between gpt-5.5 and gpt-5.6-terra? Pricing is similar, but it's hard to tell if it's better..
    • netcraft 1 hour ago
      this is exactly my question. I would expect that luna is analogous to mini before, but is terra equivalent/better than 5.5 and Sol is a step above? or is terra nerfed and 5.5 is analogous to sol?
  • WarmWash 2 hours ago
    8% on ARC-AGI-3, they actually got some traction going...
    • eugene3306 2 hours ago
      note that ARC-AGI-3 has its rules changed.

      before today all the contestants were capped at $10k

  • celltalk 31 minutes ago
    I guess Plus accounts don't get access to Sol? Or is it because I am in Europe?
  • GodelNumbering 3 hours ago
    Dirac (https://github.com/dirac-run/dirac, https://dirac.run/) now supports gpt-5.6. This thing does now seem to be on the chatGPT/codex accounts yet.

    UPDATE: it is now available in chatGPT account also, they rolled it out

    • denysvitali 2 hours ago
      Will be there soon according to the last commits in the codex repo: https://github.com/openai/codex/pull/31684/changes

      Also, confirmed it works for me by using --model gpt-5.6-sol

    • esafak 19 minutes ago
      Does it support subagents?
    • BoorishBears 1 hour ago
      I used to pride myself on not being the "fonts too pointy, scroll too buttery" crowd! But AI has brought me full circle and now nothing removes my interest in reading even a single word on a page faster than purple gradient greeble-afflicted tailwind-slop models put out without stronger prompting/references

      That being said, maybe 5.6 can fix that!

      • GodelNumbering 1 hour ago
        Thanks, I needed to hear that lol. Yes, the site was an afterthought, core work took/takes most my focus. I will look into un-slopping the site soon.
  • CjHuber 2 hours ago
    > Instead of requiring developers to script every step or passing every tool response back through the model, Programmatic Tool Calling in the Responses API can filter large amounts of intermediate data, retain only what matters, and adapt its workflow along the way.

    this seems very interesting

  • egorfine 56 minutes ago
    Zero information on the knowledge cutoff. The model itself responds it's June 2024 which is weird given that GPT-5.5 has knowledge cutoff at August 2025.
  • AgentMasterRace 1 hour ago
    I never have have the issues most people talk about ... I feel like most were never Devs before ai and don't know what they actually need done when prompting. that on top of not utilizing good tools such as a codebase indexer, lsp and a project scaffold.
  • samuelknight 2 hours ago
    There is an issue on the page that causes the benchmark tables to get cut off. If you highlight and drag right you can see a few more models like Gemini and Claude Opus. It's also interesting that they introduced explicit caching, which is something that only Anthropic had for a long time.
  • karma_daemon 2 hours ago
    I wish model launches were like proper product releases

    it's impossible to _try_ it out on release!

    it's not on their codex subscription, or the web/mobile chatgpt interfaces, or aws bedrock, etc. I just cant find a working endpoint with the latest model after they announce

    • ssl-3 1 hour ago
      For me, minutes ago, as a Plus subscriber:

      I started up Codex CLI fresh. That version of Codex was 1.42.5. 5.6 wasn't in the models list.

      After I updated Codex to a newer version (0.144.0), 5.6-terra and -luna appeared in the models list (but not 5.6-sol).

      (It's impossible for me to know whether updating was causative or just correlative, but that's the timeline I experienced.)

    • O5vYtytb 2 hours ago
      The announcement says they're rolling it out over the next 24 hours or so. I think it's reasonable to do a slow-roll-out release for one of the most used products on the internet.
    • patapong 1 hour ago
      GPT-5.6 Terra just showed up in Codex for me.
  • guybedo 1 hour ago
    It's good to see labs taking into account the cost/task.

    Grok 4.5 is interesting because it's smart enough at great price. It seems gpt 5.6 is right there with great efficiency and great pricing.

    Working with Fable has been a great experience, but at the end of the day, if you can get only 10% of your work done because it just burns through tokens, that's not that interesting.

    I've been mostly using Opus and Fable high for planning and codex 5.5 medium for implementations. Claude is also the only model i can use for design tasks. If gpt 5.6 can finally deliver on the design side, it might be time to ditch the Claude sub and go full Gpt.

  • mchusma 2 hours ago
    Looks like a great set of models, but there are about 20 different thinking/model levels here in this family and they are very complex to pick the right one for the task

    E.g. for GeneBench Pro, it looks like you would always use GPT-5.6 Sol over Terra/Luna, its pareto optimal.

    For Agents Last Exam, you would maybe want Luna, then Terra, then Luna, then Sol as you increasingly budget for tasks.

    I feel that there may need to be a new auto mode in many of these cases. It selects the best model and thinking given a particular problem.

    Feels like it's going to have to go that way eventually, because here we have about 20 different model and thinking levels you could use, and they're not obvious which ones are right for the given use case.

  • sidcool 3 hours ago
    The claims are pretty bold. I think 5.6 may exceed Fable.
  • super256 1 hour ago
    On the tiny voids demo: does your Firefox js thread lock up as well, when you try to interact with it?

    https://openai.com/index/gpt-5-6/#a-leap-forward-in-design

    • fisher-brett 1 hour ago
      Yep, happens to me on Chrome as well
  • laichzeit0 2 hours ago
    So glad Fable limits just got reset. Thanks OpenAI.
  • philip1209 2 hours ago
    Will this run on Cerebas? I'm really looking forward to that.
    • paxys 2 hours ago
      Sam Altman confirmed during the initial limited release that Sol will run on Cerebras at 750 tok/sec.
    • apitman 1 hour ago
      This is the part I'm most excited about with the new release, though I'm concerned plebs like me may never get a chance to play with it
  • m3h 2 hours ago
    We have an official pelican on a bicycle from the OpenAI livestream:

    https://imgshare.cc/mz9xwut3

    • BrokenCogs 2 hours ago
      holy moly it's in THREE dimensions!

      AGI solved

      • SirMaster 2 hours ago
        So it's failing epically because it generated a tricycle instead of a bicycle?
        • minimaxir 2 hours ago
          The livestream presenter goofed and said the test is typically a tricycle, so I wonder if that's just a coverup.
          • paxys 2 hours ago
            The site they showed had the pelican riding a number of things, including a bicycle and another pelican (no, not in that way).
  • itvision 1 hour ago
    Weirdly, normally new ChatGPT releases are head and shoulders above anything else, but according to OpenAI's own evaluation, Anthropic's Mythos outperforms ChatGPT in quite a few benchmarks: https://openai.com/index/gpt-5-6/.

    ChatGPT 6 must be deep in the pipeline and will be released within the next few months. Maybe that's why this release is versioned 5.6, not 6.0.

  • realty_geek 3 hours ago
    Wow, the "Agents' Last Exam" graph looks unreal!
    • alimhaq 2 hours ago
      I mean the y axis is deceptive to make it seem like greater gains since it starts at 30%, when in reality the differences aren't great.

      Even worse, it's not a fair comparison: they purposefully just used "adaptive" instead of "max" for Fable.

      What about the graph looked so unreal to you?

    • therobots927 3 hours ago
      That’s because it’s bullshit
  • breatheoften 57 minutes ago
    i wish they had renamed chatgpt to codex instead of the other way around ...
  • bryceneal 3 hours ago
    I find that 5.5 gives me far fewer refusals than Anthropic models for security and reverse engineering work. I hope the same is true for 5.6.
  • luciana1u 3 hours ago
    GPT-5.6 Sol, Terra, and Luna. at this rate GPT-6 will be named after a parking lot and GPT-7 after whatever Elon names his next kid.
  • cmrdporcupine 1 hour ago
    Almost immediately ran into some the kind of gatekeeping I've heard Claude Code users complaining about with Fable. Not sure why, I just had it working on writing benchmarks for some CUDA kernels. Nothing security related:

    "This request requires additional safety checks, which can take extra time. Hang tight or retry with a faster model for a quicker response, though it may be less capable of handling complex requests."

    At least it gave me the option of waiting instead of just unceremoniously downgrading me. Appears to be making progress but... weird?

  • hughw 3 hours ago
    If it's not dangerous enough to be classified as WMD by USG, who's interested.
  • lukebuehler 2 hours ago
    Oh man, I love capitalism spoiling us here. I was just enjoying my extra Fable credits, now I'll switch to using 5.6 this weekend. I was planning to ration my Anthropic credits, I guess now I do not have to. And I was half wondering if exactly this would happen: right when Fable usage credits were starting to kick in for people, OAI swoops in and takes the puck. As much the AI craze is crazy, this play by play part is pretty fun.
    • sidrag22 2 hours ago
      top it off with anthropic stressing about the release and resetting usage to 0 for the week just now.
    • halfmatthalfcat 1 hour ago
      Anthropic just reset all limits, including Fable. Capitalism is spoiling us.
    • mempko 1 hour ago
      If hoarding is spoiling. You know what would be better than using fable and gpt 5.6, being able to run that level of model on your own hardware.
  • vinhnx 2 hours ago
  • gverrilla 21 minutes ago
    If Fable is removed from my Anthropic sub, I'll have to change to OpenAI.
  • sunaookami 1 hour ago
    Overloaded in Codex, no indication if it is already in ChatGPT and I can't use it in the API even though it says it should be available. Typical horrible OpenAI launch. Glad that Anthropic just reset the rate limits so I will go back to Fable again.
  • big_toast 3 hours ago
    The cost & output token charts are useful but I wish I could view them more like a 3D surface. Like the CS:APP memory mountain charts.

    I wonder how long model size and effort will be a few discrete points instead of continuous.

  • dwa3592 2 hours ago
    This marketing video on the page is nice!! can't wait for the hardware to get cheaper to live the AI life i wanna live.
  • neuropacabra 2 hours ago
    Is it available in EU? I only see 5.5 still :-(
    • yokoprime 1 hour ago
      Arguably not in the EU, but I'm seeing Sol,Terra and Luna on my account here in Norway
  • jstummbillig 2 hours ago
    "GPT‑5.6 is available starting today across ChatGPT, Codex, and the OpenAI API. The rollout is starting globally now and will continue gradually toward full availability over the next 24 hours."
    • terramex 2 hours ago
      I am on Plus subscription and see Terra and Luna in Codex, but no sign of Sol. Will it be available only on Pro plans?
      • lostmsu 2 hours ago
        I an on Pro and it still returns "The 'gpt-5.6' model is not supported when using Codex with a ChatGPT account"

        UPD from announcement: "The rollout is starting globally now and will continue gradually toward full availability over the next 24 hours."

    • enraged_camel 2 hours ago
      My Codex app got upgraded to the new unified ChatGPT app. I don't see Sol available though. Only Terra and Luna. I'm on the Pro plan. Anyone else see it?
      • leonidasv 42 minutes ago
        Same here (Business Plan).
      • RobinL 2 hours ago
        Same, no Sol (i'm on plus)
  • vamsiraju 2 hours ago
    prompts -> loops -> slingshots?

    Its an extremely capable model. I think the way we need to approach works shifts again. We need to get our harnesses/workflows to let it gather some momentum on the first couple rounds but then we also need to structure it so that it can slingshot and accomplish the long range goal.

  • artisin 3 hours ago
    I wish they had kept their previous sensible naming convention instead of this celestial Sol, Terra, and Luna mumbo-jumbo
    • SwellJoe 13 minutes ago
      I assume they're jealous of the Fable/Mythos hype. People talk about Fable like it's a whole new thing, rather than another incremental improvement over the existing best models (which has happened several times and continues to happen).
  • vatsachak 2 hours ago
    I wonder what increment of progress will be achieved by the next billion dollars
  • kubb 3 hours ago
    They have a fantastic media team.
  • hereme888 2 hours ago
    So is 5.5-Daybreak still relevant for cyber security give. 5.6 capabilities?
  • fractorial 2 hours ago
    Sounds like a perfect fit for a minimal or bespoke harness?
  • newfriend 3 hours ago
    >Even at medium reasoning, it beats Fable 5 by 11.4 points at roughly one-quarter the estimated cost.

    Sounds great.

    Also latency looks very good.

  • ls612 1 hour ago
    I think the most interesting part of this is that OpenAI is going way easier on the classifiers than Anthropic. They explicitly state that many defensive cybersecurity uses are supported and implicitly criticize Anthropic's stance on Fable's uses by saying that overblocking cyber requests is itself a major security risk as more AI models continue to advance in intelligence. I have so many questions as to what is going on on a game theoretic level in the AI space in the past two months, it seems like multiple actors have realized their incentives are really quite different than they originally thought.
  • hyperknot 2 hours ago
    > GPT‑5.6 also introduces more predictable prompt caching, including support for explicit cache breakpoints (opens in a new window) and a 30-minute minimum cache life.

    Great to read they are moving away from the 5 minute cache defaults. Hopefully other providers follow soon!

  • paul7986 1 hour ago
    For writing GPT which i was subscribed to Fall 2024 to March 2026 (laid off) is superior to Gemini. Been using Gemini since March mostly and they offered a $10 a month plan so i took it. Though today realizing GPT is superior to help me write I am back to being a paying customer. Im in full swing mode to get back into the job market (get the heck away from UI/UX which is now a stupid career in terms of number of jobs out there and in the future there will continue to be less) pivoting into product management (can vibe code anything now) and or customer relations. Hopefully GPT helps me with this pivot and Im again gainfully employed!
  • EugeneOZ 1 hour ago
    GPT 5.6 Sol is a token hog. After implementing the task, it started some "reviews" I didn't ask for - they consumed 19.5M and 11.9M tokens, while the task itself was below 5M tokens.
  • mydreamof 53 minutes ago
    Bro these colors on chars are unbelievlable, I can not understand which is opus, which is fable, which is GPT...
  • maxdo 2 hours ago
    cursor benchmarks with GPT 5.6 in picture, a good reason to stop using opus.

    https://cursor.com/evals

    The good news you don't have to send your dollars to China to fund ai dictatorship, in russia, north korea, african countries and south america.

  • ddxv 2 hours ago
    I'm disappointed these models continue to be closed source and so expensive.

    Open weight models being 10x or more cheaper is just so much more of an unlock than incremental gains for me.

    • AgentMasterRace 36 minutes ago
      because they're stealing from the frontier models. they're gaming the benchmarks. look how bad glm 5.2 is on cursors evals. gmhit garbage , but it gets glazed as God tier.
  • nharziro 3 hours ago
    where is it? Still not accessible...
    • jstummbillig 2 hours ago
      "GPT‑5.6 is available starting today across ChatGPT, Codex, and the OpenAI API. The rollout is starting globally now and will continue gradually toward full availability over the next 24 hours."
  • brcmthrowaway 1 hour ago
    Benchmaxxed
  • simianwords 2 hours ago
    > On the Artificial Analysis Coding Agent Index, GPT‑5.6 Sol with max reasoning sets a new state of the art at 80, 2.8 points above Fable 5, while using less than half the output tokens, taking less than half the time, and costing about one-third less.

    > That advantage extends across the family: Terra performs just above Fable 5, while Luna outperforms Opus 4.8; each does so in roughly one-third of the time, with about half as many output tokens, and at approximately one-quarter the estimated cost.

    Wow. I don't believe it. Every indication and twitter post told me that Fable is much more intelligent than Sol and here we are told that even Terra outperforms Fable?

    Not only that, Sol doesn't even come with run time classifiers. So it is even more suspicious.

    What's even stranger is that OpenAI is directly referencing a competitor in this direct way.

  • saberience 3 hours ago
    "On Agents’ Last Exam (opens in a new window), an evaluation of long-running professional workflows across 55 fields, GPT‑5.6 Sol sets a new high of 53.6, eclipsing Claude Fable 5 (adaptive reasoning) by 13.1 points. Even at medium reasoning, it beats Fable 5 by 11.4 points at roughly one-quarter the estimated cost. That efficiency extends to smaller models, which are essential to making intelligence more abundant and affordable: GPT‑5.6 Terra and GPT‑5.6 Luna outperform Fable 5 at around one-sixteenth the cost. "

    Some pretty big claims and results! Excited to see how it feels during usage.

    I use Fable and 5.5 extensively and I still find both have a place in my toolkit, i.e. Fable IS good but it isn't perfect, and it's still better to play them off against each other. I have Fable and 5.5 write plans and have them adversarially review each other's plans.

    Having this amount of competition in the coding model space is good for all of us.

    • vamsiraju 1 hour ago
      I think this is the phase shift 5.6 (Sol set to Ultra) is bringing to the table. Until now we have become accustomed to asking models to continue and their natural inclination is always to stop. Now OpenAI have flipped it around and for the first time are asking us to steer or stop the model instead, and its own inclination is to keep going. We now have to decide when we need to steer or want to catch up on our understanding of the work done but it will keep going.
  • golangdev 1 hour ago
    good alternative to anthropic
  • diwank 1 hour ago
    i'm not happy with how openai is trying to pit 5.6 sol as a cheaper equivalent to fable here

    for one thing, they said that on AA, sol is "within one point of fable" at 58.9 vs 59.9 but don't clarify that the latter is with safeguards where ~8% of the tasks got routed to opus

    i'm not rooting for either and genuinely think that the token efficiency and cheaper price are important but this sort of thing just feels disingenuous :-/

    • mnicky 48 minutes ago
      This is especially interesting because IIRC the AA benchmark is calibrated so that 1 point and greater difference is statistically significant.
  • browski 2 hours ago
    Here's me using a Gemini chat log scraper (from Gdrive) then dumping my prompt+Gemini response into local AI

    Never go over the free limits in Gemini Pro.

    Gemini is great at research and architecture, and my 30 years experience in programming everything; for fun or work; means together there is little to no code slop.

    Add to project repo some git submodules of reference source code; boom, bobs your uncle

    Zero reason to sign up for OAI or Claude. With employers realizing the costs are more than employees, local models getting more powerful, and models in chips just a few years out, neither of the one note LLM companies without diversified services and R&D portfolios gonna last

  • OutOfHere 2 hours ago
    Like the last time, again they failed to note whether there is an Instant model or when it might become available.
  • Razengan 2 hours ago
    Just a day before my $100 subscription expires, perfect
  • therobots927 3 hours ago
    Do they expect us this model is 15ppt more accurate at half the price of fable? What’s going on?
  • dude250711 3 hours ago
    Not available - checked and it's not there.
    • tedsanders 2 hours ago
      As usual, even though GPT-5.6 is releasing today, the rollout in ChatGPT and Codex will be gradual over many hours so that we can make sure service remains stable for everyone (same as our previous launches). We usually start with Pro/Enterprise accounts and then work our way down to Plus. We know it's slightly annoying to have to wait a random amount of time, but we do it this way to keep service maximally stable.

      The timescale is typically hours not minutes, so if you don't see it now, I'd try again later today.

      We mention it will be a gradual rollout over the next 24 hours in the Availability section at the bottom of the blog but I admit it's pretty buried.

      (I work at OpenAI.)

    • efficax 2 hours ago
      on Plus I see Terra and Luna, but not Sol
    • trollbridge 2 hours ago
      It's available in Cursor now.
    • cactusplant7374 3 hours ago
      They have really been stringing us along for the past few weeks.
  • bearmania 2 hours ago
    If OpenAI can add all the features from CC into Codex i’ll gladly switch.
    • bel8 2 hours ago
      Which features?
    • OutOfHere 2 hours ago
      You posted the same comment twice.
  • tipiirai 3 hours ago
    Thought Fable was great
  • willchis 3 hours ago
    The marketing team must've done research that said "people are starting to think that you guys are evil-water-stealing-lay-off-loving-bubble-bursting scumbags" and decided to really lean into the small family business and happy font vibes!
  • bearmania 2 hours ago
    if OpenAI adds all the features from CC into Codex, i’ll gladly switch.
    • gavino 2 hours ago
      What features are you missing? That you can't add through skills?
  • gozucito 3 hours ago
    The meat of the report for SWEs:

    SWE-Bench Pro Sol: 64.6% Fable: 80% Opus: 69.2% (!!!!)

    So, it still trails Opus, significantly, and is not a next-gen coding model like Mythos/Fable 5.

    Disappointing to say the least, but somewhat expected.

    • osti 2 hours ago
      SWE-Bench pro is pretty much useless now even though many ppl still look at it. OpenAI published a report yesterday saying so as well. Only look at DeepSWE and FrontierCode right now for coding imo.
      • SirMaster 2 hours ago
        Amazing, a company that does poorly in a benchmark says that benchmark is useless...
        • osti 2 hours ago
          SWE-bench series just aren't that great by today's standard, even Anthropic previously stated Claude had memorized solutions for the non Pro version of the benchmark, I suspect the recent increase in the score for the Pro version probably also had similar behaviors.

          But anyway, I think it's pretty useless to look at SWE Bench's now when other way better benchmarks exist.

          • tripleee 1 hour ago
            > Anthropic previously stated Claude had memorized solutions for the non Pro version of the benchmark

            yeah that was the point of introducing the Pro version

    • sk4rekr0w 2 hours ago
      You've overstated the conclusion. The SWE-bench series has had issues since its inception.

      OpenAI no longer recommends SWE-Bench-Pro as a benchmark: https://openai.com/index/separating-signal-from-noise-coding...

      • bigfishrunning 28 minutes ago
        That's smart, they only recommend benchmarks that make them look better then their competitors.
    • paxys 3 hours ago
      Makes sense why they released an entire study yesterday discrediting SWE-bench Pro.
      • osti 2 hours ago
        And they'd be right, it's an almost saturated benchmark where even some subpar open source models score very well on. And most models are clustered within a small range so it really doesn't tell you much.
    • CuriouslyC 2 hours ago
      [dead]
  • system2 3 hours ago
    At this point, they are just changing the decimals to stay relevant and in the news.
    • dude250711 3 hours ago
      Anthropic should be grateful OpenAI did not borrow "Epic" and "Legend".
      • system2 2 hours ago
        I expect OpenAI names to be "fabulous", "glorious", "empowered", "delicious" etc.
  • infraredshift 2 hours ago
    [dead]
  • m3h 2 hours ago
    [dead]
  • fgeytk2 3 hours ago
    [dead]
  • poolnoodle 3 hours ago
    [flagged]
    • minimaxir 3 hours ago
      Then why comment?
      • Mistletoe 3 hours ago
        Apathy is a valid feeling to these incremental improvements touted as vital and amazing and worth paying for.
  • enraged_camel 3 hours ago
    CTRL-F: Fable

    15 hits

    Holy shit. They must be feeling very threatened by Fable if they're spending this much energy talking about it in the release notes for their own model.

    • cbg0 3 hours ago
      In the past they received a lot of hate for not comparing to the competition.
    • BrokenCogs 3 hours ago
      yikes - looks like you need to go back to stats school

      gemini - 13 hits

      opus - 18 hits

      So they are more threatened by opus than fable, or are they almost as threatened by gemini as they are by fable?

      • enraged_camel 2 hours ago
        The second paragraph has four mentions of Fable. I think that makes my case pretty clearly.
    • Laurel1234 2 hours ago
      Anthropic fumble of Fable's release will go down in the history books, makes sense for OpenAI to run with it.
    • therobots927 3 hours ago
      Apparently it significant outperforms fable on both an intelligence and cost index.

      I don’t believe it at all and I don’t think anyone else does either.

      • trollbridge 2 hours ago
        I believe that it outperformed it on benchmarks.
    • simianwords 2 hours ago
      Downvoted comment but I did find this comparison aggressive and tacky.
  • rvz 3 hours ago
    Most importantly, the cost:

    > GPT‑5.6 is priced per 1M tokens across three model sizes: Sol is $5 input / $30 output; Terra is $2.50 input / $15 output; and Luna is $1 input / $6 output.

    Just as expensive as Fable 5. But of course, another slot machine upgrade but the costs will keep going up and the open weight models from china will continue to race everyone else to $0.

    Looking forward to the next version of GLM, Qwen, Deepseek and Minimax.

    • celesian 3 hours ago
      That's wrong. GPT 5.6 Sol looks to have the same price as GPT 5.5, apart from a new pricing fee for cache writes. Fable 5 is $10 input / $50 output.
    • therobots927 3 hours ago
      Also watching deepseek closely. Seems like US frontier labs only know how to throw money at things as opposed to actually make smart improvements to the algorithms.
      • trollbridge 2 hours ago
        To be fair, DeepSeek doubled prices during the peak Chinese workday. (Which admittedly doesn't affect me much.)
  • I_am_tiberius 2 hours ago
    The way they talk about cyber security fixes makes clear that they are in bed with the government in order to get ahead of Anthropic.
    • culi 2 hours ago
      All of them closely collaborate with the government. LLMs are a national security priority and are vetted. Claude AI was used by Palantir's Maven to target the Minab school that led to a triple tap strike killing over 150 schoolchildren.
      • applicative 1 hour ago
        The Minab disaster has every sign of being a pure humint fail the defense department decided to cover up with politically expedient AI blaming.