[HN Gopher] GPT-5.3-Codex
___________________________________________________________________
GPT-5.3-Codex
Author : meetpateltech
Score : 872 points
Date : 2026-02-05 18:08 UTC (4 hours ago)
(HTM) web link (openai.com)
(TXT) w3m dump (openai.com)
| minimaxir wrote:
| I remember when AI labs coordinated so they didn't push major
| announcements on the same day to avoid cannibalizing each other.
| Now we have AI labs pushing major announcements _within 30
| minutes_.
| crorella wrote:
| The thrill of competition
| zozbot234 wrote:
| They're also coordinating around Chinese New Year to compete
| with new releases of the major open/local models.
| DonHopkins wrote:
| Year of the Pelican!
| hoeoek wrote:
| simonw?
| IhateAI wrote:
| A sign of the inevitible implosion !
| observationist wrote:
| The labs have fully embraced the cutthroat competition, the
| arms race has fully shed the civilized facade of beneficient
| mutual cooperation.
|
| Dirty tricks and underhanded tactics will happen - I think
| Demis isn't savvy in this domain, but might end up stomping out
| the competition on pure performance.
|
| Elon, Sam, and Dario know how to fight ugly and do the nasty
| political boardroom crap. 26 is gonna be a very dramatic year,
| lots of cinematic potential for the eventual AI biopics.
| manquer wrote:
| >civilized facade of mutual cooperation
|
| >Dirty tricks and underhanded tactics
|
| As long the tactics are legal ( i.e. not corporate espionage,
| bribes etc), the no holds barred full free market competition
| is the best thing for the market and the consumers.
| thethimble wrote:
| The consumers are getting huge wins.
|
| Model costs continue to collapse while capability improves.
|
| Competition is fantastic.
| mrandish wrote:
| > The consumers are getting huge wins.
|
| However, the investors currently subsidizing those wins
| to below cost may be getting huge losses.
| ajam1507 wrote:
| > As long the tactics are legal ( i.e. not corporate
| espionage, bribes etc), the no holds barred full free
| market competition is the best thing for the market and the
| consumers.
|
| The implicit assumption here is that we have constructed
| our laws so skillfully that the only path to win a free
| market competition is by producing a better product, or
| that all efforts will be spent doing so. This is never the
| case. It should be self-evident from this that there is a
| more productive way for companies to compete and our laws
| are not sufficient to create the conditions.
| dwaltrip wrote:
| Sure, it can be beneficial. But don't forget that
| externalities are a thing.
| wiz21c wrote:
| in the short term maybe, in the long term it depends how
| many winners you have. If only two, the market will be a
| duopoly. Customers will get better AI but will have zero
| power over the way the AI is produced or consumed (i.e. cO2
| emission, ethics, etc will be burnt)
| tedsanders wrote:
| This goes way back. When OpenAI launched GPT-4 in 2023, both
| Anthropic and Google lined up counter launches (Claude and
| Magic Wand) right before OpenAI's standard 10am launch time.
| manquer wrote:
| Wouldn't that be illegal ? i.e. cartel to collude like that ?
| cedws wrote:
| I wish they'd just stop pretending to care about safety, other
| than a few researchers at the top they care about safety only
| as long as they aren't losing ground to the competition. Game
| theory guarantees the AI labs will do what it takes to ensure
| survival. Only regulation can enforce the limits, self policing
| won't work when money is involved.
| thethimble wrote:
| As long as China continues to blitz forward, regulation is a
| direct path to losing.
| pixl97 wrote:
| You mean all paths are direct paths to losing.
| cedws wrote:
| Define "losing."
|
| Europe is prematurely regarded as having lost the AI race.
| And yet a large portion of Europe live higher quality lives
| compared to their American counterparts, live longer, and
| don't have to worry about an elected orange unleashing
| brutality on them.
| thethimble wrote:
| If the world is built on AI infrastructure (models,
| compute, etc.) that is controlled by the CCP then the
| west has effectively lost.
|
| This may lead to better life outcomes, but if the west
| doesn't control the whole stack then they have lost their
| sovereignty.
|
| This is already playing out today as Europe is dependent
| on the US for critical tech infrastructure (cloud, mail,
| messaging, social media, AI, etc). There's no home grown
| European alternatives because Europe has failed to create
| an economic environment to assure its technical
| sovereignty.
| fakedang wrote:
| Europe has already lost the tech race - their cloud
| systems that their entire welfare states rely upon are
| all hosted on servers hosted by American private
| companies, which can turn them off with a flick of a
| switch if and when needed.
|
| When the welfare state, enabled by technology, falls
| apart, it won't take long for European society to fall
| apart. Except France maybe.
| vovavili wrote:
| The last thing I would want is for excessively neurotic
| bureaucrats to interfere with all the mind-blowing progress
| we've had in the last couple of years with LLM technology.
| granzymes wrote:
| I think Anthropic rushed out the release before 10am this morning
| to avoid having to put in comparisons to GPT-5.3-codex!
|
| The new Opus 4.6 scores 65.4 on Terminal-Bench 2.0, up from 64.7
| from GPT-5.2-codex.
|
| GPT-5.3-codex scores 77.3.
| __jl__ wrote:
| Impressive jump for GPT-5.3-codex and crazy to see two top
| coding models come out on the same day...
| granzymes wrote:
| Insane! I think this has to be the shortest-lived SOTA for
| any model so far. Competition is amazing.
| the_duke wrote:
| I do not trust the AI benchmarks much, they often do not line
| up with my experience.
|
| That said ... I do think Codex 5.2 was the best coding model
| for more complex tasks, albeit quite slow.
|
| So very much looking forward to trying out 5.3.
| NitpickLawyer wrote:
| Just some anecdata++ here but I found 5.2 to be really good
| at code review. So I can have something crunched by cheaper
| models, reviewed async by codex and then re-prompt with the
| findings from the review. It finds good things, doesn't flag
| nits (if prompted not to) and the overall flow is worth it
| for me. Speed loss doesn't impact this flow that much.
| kilroy123 wrote:
| Personally, I have Claude do the coding. Then 5.2-high do
| the reviewing.
| seunosewa wrote:
| Then I pass the review back to Claude Opus to implement
| it.
| VladVladikoff wrote:
| Just curious is this a manual process or you guys have
| automated these steps?
| ricketycricket wrote:
| I have a `codex-review` skill with a shell script that
| uses the Codex CLI with a prompt. It tells Claude to use
| Codex as a review partner and to push back if it
| disagrees. They will go through 3 or 4 back-and-forth
| iterations some times before they find consensus. It's
| not perfect, but it does help because Claude will point
| out the things Codex found and give it credit.
| bryanlarsen wrote:
| Mind sharing the skill/prompt?
| _zoltan_ wrote:
| zen-mcp (now called pal-mcp I think) and then claude code
| can actually just pass things to gemini (or any other
| model)
| kilroy123 wrote:
| Sometimes, depends on how big of a task. I just find 5.2
| so slow.
| _zoltan_ wrote:
| I have Opus 4.5 do everything then review it with Gemini
| 3.
| StephenHerlihyy wrote:
| I don't use OpenAI too much, but I follow a similar work
| flow. Use Opus for design/architecture work. Move it to
| Sonnet for implementation and build out. Then finally over
| to Gemini for review, QC and standards check. There is an
| absolute gain in using different models. Each has their own
| style and way of solving the problem just like a human
| team. It's kind of awesome and crazy and a bit scary all at
| once.
| readyforbrunch wrote:
| How do you orchestrate this workflow? Do you define
| different skills that all use different models, or
| something else?
| jahsome wrote:
| Another day, another hn thread of "this model changes
| everything" followed immediately by a reply stating "actually
| I have the literal opposite experience and find competitor's
| model is the best" repeated until it's time to start the next
| day's thread.
| malshe wrote:
| This pretty accurately summarizes all the long discussions
| about AI models on HN.
| clhodapp wrote:
| And of course the benchmarks are from the school of "It's
| better to have a bad metric than no metric", so there
| really isn't any way to falsify anyone's opinions...
| cactusplant7374 wrote:
| Hourly occurrence on /r/codex. Model astrology is about the
| vibes.
| StephenHerlihyy wrote:
| What amazes me the most is the speed at which things are
| advancing. Go back a year or even a year before that and
| all these incremental improvements have compounded. Things
| that used to require real effort to consistently solve,
| either with RAGs, context/prompt engineering, have
| become... trivial. I totally agree with your point that
| each step along the way doesn't necessarily change that
| much. But in the aggregate it's sort of insane how fast
| everything is moving.
| Rudybega wrote:
| The denial of this overall trend on here and in other
| internet spaces is starting to really bother me. People
| need to have sober conversations about the speed of this
| increase and what kind of effects it's going to have on
| the world.
| SatvikBeri wrote:
| I use Claude Code every day, and I'm not certain I could
| tell the difference between Opus 4.5 and Opus 4.0 if you
| gave me a blind test
| fooker wrote:
| Yeah, these benchmarks are bogus.
|
| Every new model overfits to the latest overhyped benchmark.
|
| Someone should take this to a logical extreme and train a
| tiny model that scores better on a specific benchmark.
| mrandish wrote:
| > Yeah, these benchmarks are bogus.
|
| It's not just over-fitting to leading benchmarks, there's
| also too many degrees of freedom in how a model is tested
| (harness, etc). Until there's standardized documentation
| enabling independent replication, it's all just
| benchmarketing .
| fooker wrote:
| For the current state of AI, the harness is unfortunately
| part of the secret sauce.
| scoring1774 wrote:
| This has been done: https://arxiv.org/abs/2510.04871v1
| bunderbunder wrote:
| All shared machine learning benchmarks are a little bit
| bogus, for a really "machine learning 101" reason: your
| test set only yields an unbiased performance metric if you
| agree to only use it once. But that just isn't a realistic
| way to use a shared benchmark. Using them repeatedly is
| kind of the whole point.
|
| But even an imperfect yardstick is better than no yardstick
| at all. You've just got to remember to maintain a healthy
| level of skepticism is all.
| abustamam wrote:
| Is an imperfect yardstick better than no yardstick? It
| reminds me of documentation -- the only thing worse than
| no documentation is wrong documentation.
| aurareturn wrote:
| 5.2 Codex became my default coding model. It "feels" smarter
| than Opus 4.5.
|
| I use 5.2 Codex for the entire task, then ask Opus 4.5 at the
| end to double check the work. It's nice to have another
| frontier model's opinion and ask it to spot any potential
| issues.
|
| Looking forward to trying 5.3.
| koakuma-chan wrote:
| Opus 4.5 is more creative and better at making UIs
| nerdsniper wrote:
| Opus 4.5 still worked better for most of my work, which is
| generally "weird stuff". A lot of my programming involves
| concepts that are a bit brain-melting for LLMs, because
| multiple "99% of the time, assumption X is correct" are
| reversed for my project. I think Opus does better at not
| falling into those traps. Excited to try out 5.3
| nubg wrote:
| what do you do?
| nurettin wrote:
| Opus was quite useless today. Created lots of globals, statics,
| forward declarations, hidden implementations in cpp files with
| no testable interface, erasing types, casting void pointers, I
| had to fix quite a lot and decouple the entangled mess.
|
| Hopefully performance will pick up after the rollout.
| nickstinemates wrote:
| Did you give it any architecture guidance? An architecture
| skill that it can load to make sure it lays out things
| according to your taste?
| nurettin wrote:
| Yes, it has a very tight CLAUDE.md which it used to follow.
| Feels like this happens a couple of times a month.
| leumon wrote:
| they tested it at xhigh reasoning though, which is probably
| double the cost of Anthropic's model.
|
| Cost to Run Artificial Analysis Intelligence Index:
|
| GPT-5.2 Codex (xhigh): $3244
|
| Claude Opus 4.5-reasoning: $1485
|
| (and probably similar values for the newer models?)
| redox99 wrote:
| With $20 gpt plan you can use xhigh no problem. With $20
| Claude plan you reach the 5h limit with a single feature.
| mattkevan wrote:
| Ha, Claude Code on a pro plan often can't complete a single
| message before hitting the 5h limit. Not hit it once so far
| on Codex.
| naths88 wrote:
| This, so frustrating. But CC is so much faster too.
| Computer0 wrote:
| A provider's API costs seemingly do not reflect each
| respective SOTA provider's subscription usage allowances.
| wilg wrote:
| In my personal experience the GPT models have always been
| significantly better than the Claude models for agentic coding,
| I'm baffled why people think Claude has the edge on
| programming.
| dudeinhawaii wrote:
| I think for many/most programmers = 'speed + output' and
| webdev == "great coding".
|
| Not throwing shade anyone's way. I actually do prefer Claude
| for webdev (even if it does cringe things like generate
| custom CSS on every page) -- because I hate webdev and Claude
| designs are always better looking.
|
| But the meat of my code is backend and "hard" and for that
| Codex is always better, not even a competition. In that
| domain, I want accuracy and not speed.
|
| Solution, use both as needed!
| whynotminot wrote:
| > Solution, use both as needed!
|
| This is the way. People are unfortunately starting to
| divide themselves into camps on this -- it's human nature
| we're tribal - but we should try to avoid turning this into
| a Yankees Redsox.
|
| Both companies are producing incredible models and I'm glad
| they have strengths because if you use them both where
| appropriate it means you have more coverage for important
| work.
| falloutx wrote:
| > I actually do prefer Claude for webdev
|
| Ah and let me guess all your frontends look like cookie
| cutter versions of this: https://openclaw.dog/
| Yiin wrote:
| Yes and I love it.
| soulofmischief wrote:
| GPT 5.2 codex plans well but fucks off a lot, goes in circles
| (more than opus 4.5) and really just lacks the breadth of
| integrated knowledge that makes opus feel so powerful.
|
| Opus is the first model I can trust to just do things, and do
| them right, at least small things. For larger/more complex
| things I have to keep either model on extremely short
| leashes. But the difference is enough that I canceled my GPT
| Pro sub so I could switch to Claude. Maybe 5.3 will change
| things, but I also cannot continue to ethically support Sam
| Altman's business.
| wilg wrote:
| I always use 5.2-Codex-High or 5.2-Codex-Extra High (in
| Cursor). The regular version is probably too dumb.
| fragmede wrote:
| How many people are building the same thing multiple times to
| compare model performance? I'm much more interested in
| getting the thing I'm building getting built, than than
| comparing AIs to each other.
| jronak wrote:
| Did you look at the ARC AGI 2? Codex might be overfit for
| terminal bench
| tedsanders wrote:
| ARC AGI 2 has a training set that model providers can choose
| to train on, so really wouldn't recommend using it as a
| general measure of coding ability.
| janalsncm wrote:
| More fundamentally, ARC is for abstract reasoning. Moving
| blocks around on a grid. While in theory there is some
| overlap with SWE tasks, what I really care about is
| competence on the specific task I will ask it to do. That
| requires a lot of domain knowledge.
|
| As an analogy, Terence Tao may be one of the smartest
| people alive now, but IQ alone isn't enough to do a job
| with no domain-specific training.
| mrandish wrote:
| A key aspect of ARC AGI is to remain highly resistant to
| training on test problems which is essential for ARC AGI's
| purpose of evaluating fluid intelligence and adaptability
| in solving novel problems. They do release public test sets
| but hold back private sets. The whole idea is being a test
| where training on public test sets doesn't materially help.
|
| The only valid ARC AGI results are from tests done by the
| ARC AGI non-profit using an unreleased private set. I
| believe lab-conducted ARC AGI tests must be on public sets
| and taken on a 'scout's honor' basis that the lab self-
| administered the test correctly, didn't cheat or
| accidentally have public ARC AGI test data slip into their
| training data. IIRC, some time ago there was an issue when
| OpenAI published ARC AGI 1 test results on a new model's
| release which the ARC AGI non-profit was unable to
| replicate on a private set some weeks later (to be fair, I
| don't know if these issues were resolved). _Edit to Add:_
| Summary of what happened: https://grok.com/share/c2hhcmQtMw
| _66c34055-740f-43a3-a63c-4b...
|
| I have no expertise to verify how training-resistant ARC
| AGI is in practice but I've read a couple of their papers
| and was impressed by how deeply they're thinking through
| these challenges. They're clearly trying to be a unique
| test which evaluates aspects of 'human-like' intelligence
| other tests don't. It's also not a specific coding test and
| I don't know how directly ARC AGI scores map to coding
| ability.
| fishpham wrote:
| Model card:
| https://cdn.openai.com/pdf/23eca107-a9b1-4d2c-b156-7deb4fbc6...
| raincole wrote:
| Almost like Anthropic and OpenAI are trying to front run each
| other
| kingstnap wrote:
| That was fast!
|
| I really do wonder whats the chain here. Did Sam see the Opus
| announcement and DM someone a minute later?
| Mond_ wrote:
| OpenAI has a whole history of trying to scoop other providers.
| This was a whole thing for Google launches, where OpenAI
| regularly launched something just before Google to grab the
| media attention.
| rsanek wrote:
| Some recent examples:
|
| GPT-4o vs. Google I/O (May 2024): OpenAI scheduled its
| "Spring Update" exactly 24 hours before Google's biggest
| event of the year, Google I/O. They launched GPT-4o voice
| mode.
|
| Sora vs. Gemini 1.5 Pro (Feb 2024): Just two hours after
| Google announced its breakthrough Gemini 1.5 Pro model, Sam
| Altman tweeted the reveal of Sora (text-to-video).
|
| ChatGPT Enterprise vs. Google Cloud Next (Aug 2023): As
| Google began its major conference focused on selling AI to
| businesses, OpenAI announced ChatGPT Enterprise.
| maxpert wrote:
| Tell me that you are hurt without telling me that you are hurt
| this applies to Sam right now
| NewsaHackO wrote:
| I assume some sort of corporate espionage. This is high stakes
| after all
| maheshrijal wrote:
| It seems Fast!
| simianwords wrote:
| Any notes on pricing?
| Tiberium wrote:
| It's not in the API yet - "We are working to safely enable API
| access soon.", but I assume the rate-limits won't be worse than
| for 5.2 Codex.
| nine_k wrote:
| Ah, "It's ready, but not yet".
| yunyu wrote:
| You can just use it outside of the API?
| kingstnap wrote:
| > GPT-5.3-Codex was co-designed for, trained with, and served on
| NVIDIA GB200 NVL72 systems. We are grateful to NVIDIA for their
| partnership.
|
| This is hilarious lol
| uh_uh wrote:
| How so?
| Philpax wrote:
| They're on shaky ground right now
| https://arstechnica.com/information-
| technology/2026/02/five-...
| kingstnap wrote:
| Its kind of a suck up that more or less confirms the beef
| stories that were floating around this past week.
|
| In case you missed it. For example:
|
| Nvidia's $100 billion OpenAI deal has seemingly vanished -
| Ars Technica
|
| https://arstechnica.com/information-
| technology/2026/02/five-...
|
| Specifically this paragraph is what I find hilarious.
|
| > According to the report, the issue became apparent in
| OpenAI's Codex, an AI code-generation tool. OpenAI staff
| reportedly attributed some of Codex's performance limitations
| to Nvidia's GPU-based hardware.
| esafak wrote:
| > OpenAI staff reportedly attributed some of Codex's
| performance limitations to Nvidia's GPU-based hardware.
|
| They should design their own hardware, then. Somehow the
| other companies seem to be able to produce fast-enough
| models.
| dajonker wrote:
| There was never a $100 billion deal. Only a letter of
| intent which doesn't mean anything contractually.
| shibeprime wrote:
| I know we just got a reset and a 2x bump with the native app
| release, but shipping 5.3 with no reset feels mismatched. If I'd
| known this was coming, I wouldn't have used up the quota on the
| previous model.
| trilogic wrote:
| When 2 multi billion giants advertise same day, it is not
| competition but rather a sign of struggle and survival. With all
| the power of the "best artificial intelligence" at your
| disposition, and a lot of capital also all the brilliant minds,
| THIS IS WHAT YOU COULD COME UP WITH?
|
| Interesting
| rishabhaiover wrote:
| What happened to you?
| raincole wrote:
| AI fried brains, unfortunately.
| wasmainiac wrote:
| I mean, he has a point it's just not very eloquently
| written.
| trilogic wrote:
| I empathize with the situation, no elegance from them, no
| eloquence from me :)
| sdf2erf wrote:
| Yeah they are both fighting for survival. No surprise really.
|
| Need to keep the hype going if they are both IPO'ing later this
| year.
| superze wrote:
| How many IPOs can a company really do?
| re-thc wrote:
| As many as they want. They can "spin off" and then "merge"
| again.
| thethimble wrote:
| The AI market is an infinite sum market.
|
| Consider the fact that 7 year old TPUs are still sitting at
| near 100p utilization today.
| lossolo wrote:
| What's funny is that most of this "progress" is new datasets +
| post-training shaping the model's behavior (instruction +
| preference tuning). There is no moat besides that.
| Davidzheng wrote:
| "post-training shaping the models behavior" it seems from
| your wording that you find it not that dramatic. I rather
| find the fact that RL on novel environments providing steady
| improvements after base-model an incredibly bullish signal on
| future AI improvements. I also believe that the capability
| increase are transferring to other domains (or at least
| covers enough domains) that it represents a real rise in
| intelligence in the human sense (when measured in
| capabilities - not necessarily innate learning ability)
| CuriouslyC wrote:
| What evidence do you base your opinions on capability
| transfer on?
| WarmWash wrote:
| >There is no moat besides that.
|
| Compute.
|
| Google didn't announce $185 billion in capex to do
| cataloguing and flash cards.
| causalmodels wrote:
| Google didn't buy 30% of Anthropic to starve them of
| compute
| WarmWash wrote:
| Probably why it's selling them TPUs.
| riku_iki wrote:
| > is new datasets + post-training shaping the model's
| behavior (instruction + preference tuning). There is no moat
| besides that.
|
| sure, but acquiring/generating/creating/curating so much high
| quality data is still significant moat.
| edem wrote:
| So can I use this from Opencode? Because Anthropic started to
| enforce their TOS to kill the Opencode integration
| rs_rs_rs_rs_rs wrote:
| You can use Anthropic models in Opencode, make an api key and
| you're good to do(you can even use the in house Opencode
| router, Zen).
|
| What you can't do is pretend opencode is claude code to make
| use of that specific claude code subscription.
| tfehring wrote:
| OpenAI models in general, yes - `opencode auth login`, select
| OpenAI, then ChatGPT Pro/Plus. I just checked and 5.3-codex
| isn't available in opencode yet, but I assume it will be soon.
| InsideOutSanta wrote:
| Yes, OpenAI said they'd allow usage of their subscriptions in
| opencode.
| regularfry wrote:
| I've tried opus 4.5 in opencode via the GitHub Copilot API,
| mostly to see if it works all. I don't _think_ that broke any
| terms of service? But also I haven 't checked how much more
| expensive I made it for myself over just calling them directly.
| avb wrote:
| You can also use via Opencode Zen, Github Copilot, or probably
| any number of other model providers that Opencode integrates
| with.
|
| Not sure why everyone stays focused on getting it from
| Anthropic or OpenAI directly when there are so many places to
| get access to these models and many others for the same or less
| money.
| Robin_f wrote:
| Anthropic mostly had an advantage in speed. It feels like with a
| 25% increase in speed with Codex 5.3, they are now losing that
| advantage as well.
| smith7018 wrote:
| I just asked Opus 4.6 to debug a bug in my current changes and
| it went for 20 minutes before I interrupted it. Take that as
| you will.
| bgirard wrote:
| Doesn't feel like a useful data point without more context.
| For some hard bugs I'd be thrilled to wait 30 minutes for a
| fix, for a trivial CSS fix not so much. I've spent weeks+ of
| my career fix single bugs. Context is everything.
| smith7018 wrote:
| Sure, but I've never experienced a 20 minute wait with CC
| before. It was an architectural question but it would have
| taken a couple minutes with a definitive answer on 4.5.
| maxpert wrote:
| Is this me or Sam is being absolute sore loser he is and trying
| to steal Opus thunder?
| OutOfHere wrote:
| They both are doing this to each other.
|
| BTW, loser is spelled with a single o.
| wahnfrieden wrote:
| You could also claim that Anthropic is trying to scoop OpenAI
| by launching minutes earlier, as OpenAI has done with Google in
| the past.
|
| For downvoters, you must be naive to think these companies are
| not surveilling each other through various means.
| nickthegreek wrote:
| Why is it loser? He very well could be a sore winner here.
| koakuma-chan wrote:
| OpenAI is still the only AI company that has structured
| outputs. Anthropic now supports JSON schema but you can't
| specify array length.
| wahnfrieden wrote:
| Can you elaborate what you mean - OAI structured outputs
| means JSON schema doesn't it? So are you just saying they
| both support JSON schema but Anthropic has a limitation?
| koakuma-chan wrote:
| OpenAI, in addition to JSON schema, supports "context-
| free grammar"[0], i.e. regex and lark. Anthropic also
| supports JSON schema since a few weeks ago, but they
| don't support specifying the length of JSON array, so you
| still have to worry about the model producing invalid
| output.
|
| [0]: https://platform.openai.com/docs/guides/function-
| calling#con...
|
| One thing that pisses me off is this widespread
| misunderstanding that you can just fall back to function
| calling (Anthropic's function calling accepts JSON schema
| for arguments), and that it's the same as structured
| outputs. It is not. They just dump the JSON schema into
| the context without doing the actual structured outputs.
| Vercel's AI SDK does that and it pisses me off because
| doing that only confuses the model and prefilling works
| much better.
| jiggawatts wrote:
| Google Gemini definitely has structured output.
| binsquare wrote:
| At first try it solved a problem that 5.2 couldn't previously.
|
| Seems to be slower/thinks longer.
| OutOfHere wrote:
| It is absurd to release 5.3-Codex before first releasing 5.3.
|
| Also, there is no reason for OpenAI and Anthropic to be trying to
| one-up each other's releases on the same day. It is hell for the
| reader.
| apetresc wrote:
| Why is it absurd?
| tomashubelbauer wrote:
| I agree, I was confused about where 5.3 non Codex was.
| 5.2-Codex disappointed me enough that I won't be giving 5.3
| Codex a try, but I'm looking forward to trying 5.3 non Codex
| with Pi.
| sunaookami wrote:
| GPT-5.x in general are very disappointing, the only good chat
| model was GPT-5 in the first week before they made "the
| personality warmer" and Codex in general was always kinda
| meh.
| aurareturn wrote:
| Because Claude Code is stealing the thunder so OpenAI is
| focusing on coding now.
| whizzter wrote:
| Yeah, Claude Code is what everyone is talking about these
| days and since OpenAI has always been the spending driver
| being 2nd or 3rd fiddle just isn't acceptable if they're
| gonna justify it.
| stri8ted wrote:
| That is where the money is.
| ghosty141 wrote:
| This. I think software development is the best usecase for
| AI yet. I use it almost daily at work and it's a huge help.
|
| Enterprise customers will happily pay even 100$/mo
| subscriptions and it has a clear value proposition that can
| be decently verified.
| OutOfHere wrote:
| It's where the revenue is, but it isn't going to be where
| the profit is. Developers will easily use absurdly large
| amounts of compute, costing the AI provider a lot more than
| they receive in revenue.
| modeless wrote:
| It's so difficult to compare these models because they're not
| running the same set of evals. I think literally the only eval
| variant that was reported for both Opus 4.6 and GPT-5.3-Codex is
| Terminal-Bench 2.0, with Opus 4.6 at 65.4% and GPT-5.3-Codex at
| 77.3%. None of the other evals were identical, so the numbers for
| them are not comparable.
| input_sh wrote:
| It's better on a benchmark I've never heard of!? That is
| groundbreaking, I'm switching immediately!
| modeless wrote:
| I also wasn't that familiar with it, but the Opus 4.6
| announcement leaned pretty heavily on the TerminalBench 2.0
| score to quantify how much of an improvement it was for
| coding, so it looks pretty bad for Anthropic that OpenAI beat
| them on that specific benchmark so soundly.
|
| Looking at the Opus model card I see that they also have by
| far the highest score for a single model on ARC-AGI-2. I
| wonder why they didn't advertise that.
| input_sh wrote:
| No way! Must be a coinkydink, no way OpenAI _knew_ ahead of
| time that Anthropic was gonna put a focus on that specific
| useless benchmark as opposed to all the other useless
| benchmarks!?
|
| I'm firing 10 people now instead of 5!
| rsanek wrote:
| I usually wait to see what ArtificialAnalysis says for a direct
| comparison.
| alexhans wrote:
| Isn't the best eval the one you build yourself, for your own
| use cases and value production?
|
| I encourage people to try. You can even timebox it and come up
| with some simple things that might look initially insufficient
| but that discomfort is actually a sign that there's something
| there. Very similar to moving from not having unit/integration
| tests for design or regression and starting to have them.
| xiphias2 wrote:
| ,,GPT-5.3-Codex is the first model we classify as High capability
| for cybersecurity-related tasks under our Preparedness Framework
| , and the first we've directly trained to identify software
| vulnerabilities. While we don't have definitive evidence it can
| automate cyber attacks end-to-end, we're taking a precautionary
| approach and deploying our most comprehensive cybersecurity
| safety stack to date. Our mitigations include safety training,
| automated monitoring, trusted access for advanced capabilities,
| and enforcement pipelines including threat intelligence.''
|
| While I love Codex and believe it's amazing tool, I believe their
| preparedness framework is out of date. As it is more and more
| capable of vibe coding complex apps, it's getting clear that the
| main security issues will come up by having more and more
| security critical software vibe coded.
|
| It's great to look at systems written by humans and how well
| Codex can be used against software written by humans, but it's
| getting more important to measure the opposite: how well humans
| (or their own software) are able to infiltrate complex systems
| written mostly by Codex, and get better on that scale.
|
| In simpler terms: Codex should write secure software by default.
| mrkeen wrote:
| Is "high-capability" a stronger or weaker claim than "team of
| phd-level experts"?
|
| https://www.nbcnews.com/tech/tech-news/openai-releases-chatg...
| trcf23 wrote:
| That's just classical OpenAI trying to make us believe they're
| closing on AGI... Like all << so called >> research from them
| and Anthropic about safety alignment and that their tech is so
| incredibly powerful that guardrails should be put on them.
| ActionHank wrote:
| I heard the other day that every time someone claps another
| vibe coded project embeds the api keys in the webpage.
|
| I wonder if this will continue to be the case.
| da_grift_shift wrote:
| >Our mitigations include safety training, automated monitoring,
| trusted access for advanced capabilities, and enforcement
| pipelines including threat intelligence.
|
| "We added some more ACLs and updated our regex"
| manmal wrote:
| Please no, I don't need my quick prototypes hardened against
| every perceivable threat.
| GenerWork wrote:
| I find it very, very interesting how they demoed visuals in the
| form of the "soft SaaS" website and mentioned how it can do user
| research. Codex has usually lagged behind Claude and Gemini when
| it comes to UX, so I'm curious to see if 5.3 will take the lead
| in real world use. Perhaps it'll be available in Figma Make now?
| brokencode wrote:
| I'm hoping they add better IDE integration to track active file
| and selection. That's the biggest annoyance I have in working
| with Codex.
| imasliev wrote:
| GPT-5.2-Codex was so cool at price/value rate, hope 5.3 will not
| ruin the race with claude
| I_am_tiberius wrote:
| I'd like to know if and how much illegal use of customer prompts
| are used for training.
| renewiltord wrote:
| Oh yeah that's in the "These Are The Illegal Things We Did"
| section 7.4 in the Model Card.
| xlbuttplug2 wrote:
| "But we anonymize prompts before training!"
|
| Meanwhile the prompt: Crop this photo of my passport
| wahnfrieden wrote:
| Pelican seems much worse than the Opus 4.6 one (though the
| bicycle is more accurate):
|
| https://gist.github.com/simonw/a6806ce41b4c721e240a4548ecdbe...
| morleytj wrote:
| The behind the scenes on deciding when to release these models
| has got to be pretty insanely stressful if they're coming out
| within 30 minutes-ish of each other.
| Havoc wrote:
| It's also functionally not likely without some sort of insider
| knowledge or coordination
| morleytj wrote:
| Could be, could also be situations where things are lined up
| to launch in the near future and then a mad dash happens upon
| receiving outside news of another launch happening.
|
| I suppose coincidences happen too but that just seems too
| unlikely to believe honestly. Some sort of knowledge leakage
| does seem like the most likely reason.
| meisel wrote:
| I wonder if their "5.3" was continuously being updated, with
| regenerated benchmarks with each improvement, and they just
| stayed ready to release it when claude released
| morleytj wrote:
| This seems plausible. It would be shocking if these companies
| didn't have an automated testing suite which is recomputing
| these benchmarks on a regular basis, and uploading to a
| dashboard for supervision.
|
| Given that they already pre-approved various language and
| marketing materials beforehand there's no real reason they
| couldn't just leave it lined up with a function call to go
| live once the key players make the call.
| heraldgeezer wrote:
| Anthropic and GTP 2 new models at once?
| tosh wrote:
| Terminal Bench 2.0 | Name | Score
| | |---------------------|-------| | OpenAI Codex 5.3
| | 77.3 | | Anthropic Opus 4.6 | 65.4 |
| greenfish6 wrote:
| yea but i feel like we are over the hill on benchmaxxing, many
| times a model has beaten anthropic on a specific bench, but the
| 'feel' is that it is still not as good at coding
| AstroBen wrote:
| 'feel' is no more accurate
|
| not saying there's a better way but both suck
| thethimble wrote:
| Speak for yourself. I've been insanely productive with
| Codex 5.2.
|
| With the right scaffolding these models are able to perform
| serious work at high quality levels.
| AstroBen wrote:
| ..huh?
| helloplanets wrote:
| He wasn't saying that both of the models suck, but that
| the heuristics for measuring model capability suck
| crorella wrote:
| The variety of tasks they can do and will be asked to do is
| too wide and dissimilar, it will be very hard to have a
| transversal measurement, at most we will have area specific
| consensus that model X or Y is better, it is like saying
| one person is the best coder at everything, that does not
| exist.
| pixl97 wrote:
| Yea, we're going to need benchmarks that incorporate
| series of steps of development for a particular language
| and how good each model is at it.
|
| Like can the model take your plan and ask the right
| questions where there appear to be holes.
|
| How wide of architecture and system design around your
| language does it understand.
|
| How does it choose to use algorithms available in the
| language or common libraries.
|
| How often does it hallucinate features/libraries that
| aren't there.
|
| How does it perform as context get larger.
|
| And that's for one particular language.
| tavavex wrote:
| The 'feel' of a single person is pretty meaningless, but
| when many users form a consensus over time after a model is
| released, it feels a lot more informative than a simple
| benchmark because it can shift over time as people
| individually discover the strong and weak points of what
| they're using and get better at it.
| forrestthewoods wrote:
| At the end of the day "feel" is what people rely on to pick
| which tool they use.
|
| I'd feel unscientific and broken? Sure maybe why not.
|
| But at the end of the day I'm going to choose what I see
| with my own two eyes over a number in a table.
|
| Benchmarks are a sometimes useful to. But we are in prime
| Goodharts Law Territory.
| AstroBen wrote:
| yeah, to be honest it probably doesn't matter too much. I
| think the major models are very close in capabilities
| forrestthewoods wrote:
| I don't think this is even remotely true in practice.
|
| I honestly I have no idea what benchmarks are
| benchmarking. I don't write JavaScript or do anything
| remotely webdev related.
|
| The idea that all models have very close performance
| across all domains is a moderately insane take.
|
| At any given moment the best model for my actual projects
| and my actual work varies.
|
| Quite honestly Opus 4.5 is proof that benchmarks are
| dumb. When Opus 4.5 released no one was particularly
| excited. It was better with some slightly large numbers
| but whatever. It took about a month before everyone
| realized "holy shit this is a step function improvement
| in usefulness". Benchmarks being +15% better on SWE bench
| didn't mean a damn thing.
| falloutx wrote:
| When Anthropic beats Benchmarks its somehow earned, when
| OpenAi games it, its somehow about not feeling good at
| coding.
| karmasimida wrote:
| Your feeling is not my feeling, codex is unambiguously
| smarter model for me
| xyst wrote:
| Benchmarks are useless compared to real world performance.
|
| Real world performance for these models is a disappoint.
| jdthedisciple wrote:
| Gotta love how the game demo's page title is "threejs" - I guess
| the point was to demo its vibe-coding abilities anyway, but yea..
| prng2021 wrote:
| Did they post the knowledge cutoff date somewhere
| gwd wrote:
| gpt-5.3-codex isn't available on the API yet. From TFA:
|
| > We are working to safely enable API access soon.
| ecshafer wrote:
| Funny that this and Opus 4.6 released within minutes of each
| other. Each showing similar score improvements. Each claiming to
| be revolutionary.
| __mharrison__ wrote:
| I never really used Codex (found it to slow) just 5.2, which I
| going to be an excellent model for my work. This looks like
| another step up.
|
| This week, I'm all local though, playing with opencode and
| running qwen3 coder next on my little spark machine. With the way
| these local models are progressing, I might move all my llm work
| locally.
| andix wrote:
| I think codex got much faster for smaller tasks in the last few
| months. Especially if you turn thinking down to medium.
| raffkede wrote:
| I think the slow feeling is a UI thing in codex
| __mharrison__ wrote:
| I realize my comment was unclear. I use codex the CLI all the
| time, but generally with this invocation: `codex --full-auto
| -m gpt-5.2`
|
| However, when I use the 5.2codex model, I've found it to be
| very slow and worse (hard to quantify, but I preferred
| straight-up 5.2 output).
| foft wrote:
| Having used codex a fair bit I find it really struggles with ...
| almost anything. However using the equivalent chat gpt model is
| fantastic. I guess it's a matter of focus and being provided with
| a smaller set of code to tackle.
| ffitch wrote:
| > our team was blown away > by how much Codex was able > to
| accelerate its own development
|
| they forgot to add "Can't wait to see what you do with it"
| roya51788 wrote:
| what are the benchmarks against opus 4.6?
| itay-maman wrote:
| Something that caught my eye from the announcement:
|
| > GPT-5.3-Codex is our first model that was instrumental in
| creating itself. The Codex team used early versions to debug its
| own training
|
| I'm happy to see the Codex team moving to this kind of
| dogfooding. I think this was critical for Claude Code to achieve
| its momentum.
| aurareturn wrote:
| More importantly, this is the early steps of a model self
| improving itself.
|
| Do we still think we'll have soft take off?
| reducesuffering wrote:
| This has already been going on for years. It's just that they
| were using GPT 4.5 to work on GPT 5. All this announcement
| mean is that they're confident enough in early GPT 5.3 model
| output to further refine GPT 5.3 based on initial 5.3. But
| yes, takeoff will still happen because of this recursive self
| improvement works, it's just that we're _already_ past the
| inception point.
| mirsadm wrote:
| I can't tell if this is a serious conversation anymore.
| reducesuffering wrote:
| "Best start believing in science fiction stories. You're
| in one."
|
| https://x.com/TheZvi/status/2017310187309113781
| manmal wrote:
| I guess humans were involved in all that, so how is that
| anything but tool use?
| quinncom wrote:
| Exponential growth may look like a very slow increase at
| first, but it's still exponential growth.
| janalsncm wrote:
| Sigmoids may look like exponential growth at first, until
| they saturate. Early growth alone cannot distinguish
| between them.
| gf000 wrote:
| If it's exponential growth. It may just as well be some
| slow growth and continue to be so.
| thrance wrote:
| I think the limiting factor is capital, not code. And I doubt
| GPTX is anymore competent at raising funds than the other,
| fleshy, snake oilers...
| mrandish wrote:
| > Do we still think we'll have soft take off?
|
| There's still no evidence we'll have any take off. At least
| in the "Foom!" sense of LLMs independently improving
| themselves iteratively to _substantial_ new levels being
| reliably sustained over many generations.
|
| To be clear, I think LLMs are valuable and will continue to
| significantly improve. But self-sustaining runaway positive
| feedback loops delivering exponential improvements resulting
| in leaps of tangible, real-world utility is a substantially
| different hypothesis. All the impressive and rapid
| achievements in LLMs to date can still be true while major
| elements required for Foom-ish exponential take-off are still
| missing.
| rahulyc wrote:
| Yes, but also you'll never have any early evidence of the
| Foom until the Foom itself happens.
| janalsncm wrote:
| If only General Relativity had such an ironclad defense
| of being as unfalsifiable as Foom Hypothesis is. We
| could've avoided all of the quantum physics nonsense.
| aaaalone wrote:
| I'm only saying no to keep optimistic tbh
|
| It feels crazy to just say we might see a fundamental shift
| in 5 years.
|
| But the current addition to compute and research etc. def
| goes in this direction I think.
| 8note wrote:
| making the specifications is still hard, and checking how
| well results match against specifications is still hard.
|
| i dont think the model will figure that out on its own,
| because the human in the loop is the verification method for
| saying if its doing better or not, and more importantly,
| defining _better_
| codethief wrote:
| Sounds like the researchers behind https://ai-2027.com/ haven't
| been too far off so far.
| cootsnuck wrote:
| We'll see. The first two things that they said would move
| from "emerging tech" to "currently exists" by April 2026 are:
|
| - "Someone you know has an AI boyfriend"
|
| - "Generalist agent AIs that can function as a personal
| secretary"
|
| I'd be curious how many people know someone that is sincerely
| in a relationship with an AI.
|
| And also I'd love to know anyone that has honestly replaced
| their human assistant / secretary with an AI agent. I have an
| assistant, they're much more valuable beyond rote input-
| output tasks... Also I encourage my assistant to use LLMs
| when they can be useful like for supplementing research
| tasks.
|
| Fundamentally though, I just don't think any AI agents I've
| seen can legitimately function as a personal secretary.
|
| Also they said by April 2026:
|
| > 22,000 Reliable Agent copies thinking at 13x human speed
|
| And when moving from "Dec 2025" to "Apr 2026" they switch
| "Unreliable Agent" to "Reliable Agent". So again, we'll see.
| I'm very doubtful given the whole OpenClaw mess. Nothing
| about that says "two months away from reliable".
| Rudybega wrote:
| I think they immediately corrected their median timelines
| for takeoff to 2028 upon releasing the article (I believe
| there was a math mistake or something initially), so all
| those dates can probably be bumped back a few months.
| Regardless, the trend seems fairly on track.
| zozbot234 wrote:
| > Someone you know has an AI boyfriend
|
| MyBoyfriendIsAI is a thing
|
| > Generalist agent AIs that can function as a personal
| secretary
|
| Isn't that what MoltBot/OpenClaw is all about?
|
| So far these look like successful predictions.
| ainch wrote:
| Moltbot is an attempt to do that. Would you hire it as a
| personal secretary and entrust all your personal data to
| it?
| YawningAngel wrote:
| I don't think generative AI is even close to making model
| development 50% faster
| ponyous wrote:
| I think models are smart enough for most of the stuff, these
| little incremental changes barely matter now. What I want is the
| model that is _fast_.
| derac wrote:
| This is faster if their marketing is right, it uses
| significantly less tokens. Gemini 3 flash is very good as well.
| hubraumhugo wrote:
| Anybody else not seeing it available in Codex app or CLI yet
| (with Plus)?
| haneul wrote:
| My codex CLI didn't notice version bump available, but I
| manually did pnpm add -g @openai/codex and 5.3 was there after.
| kopollo wrote:
| Where is the google?
| hsaliak wrote:
| gemini-3-flash-preview will be GA soon i hope. /s
| rustyhancock wrote:
| Anyone remember the dot-com era when you would see one provider
| claim the most miles of fibre and then later that week another
| would have the title?
| dawidg81 wrote:
| May AI not write the code for me.
|
| May I at least understand what it has "written". AI help is good
| but don't replace real programmers completely. I'm enough copy
| pasting code i don't understand. What if one day AI will fall
| down and there will be no real programmers to write the software.
| AI for help is good but I don't want AI to write whole files into
| my project. Then something may broke and I won't know what's
| broken. I've experienced it many times already. Told the AI to
| write something for me. The code was not working at all. It was
| compiling normally but the program was bugged. Or when I was
| making some bigger project with ChatGPT only, it was mostly
| working but after a longer time when I was promting more and more
| things, everything got broken.
| katspaugh wrote:
| Honest question: have you tried evolving your code architecture
| when adding features instead of just "promting more and more
| things"?
| dawidg81 wrote:
| I've tried that too but it was almost the same, chatgpt kept
| forgetting many things about the code and project structure.
| In summary AI can get problematic for me and i get with
| troubles with it. This is like one of the reasons why I still
| prefer traditional text editor for writing code like Vim over
| a "software on steroids" like VS Code and things like that...
| pixl97 wrote:
| > What if one day AI will fall down and there will be no real
| programmers to write the software.
|
| What if you want to write something very complex now that most
| people don't understand? You keep offering more money until
| someone takes the time to learn it and accomplish it, or you
| give up.
|
| I mean, there are still people that hammer out horseshoes over
| a hot fire. You can get anything you're willing to pay money
| for.
| nubg wrote:
| Sorry but companies will not hire you but instead a person who
| learned how to code with AI. Get with the times or lose.
| cheeze wrote:
| It's that Austin Powers clip of the guy slowly getting
| smooshed by the steam roller.
| dawidg81 wrote:
| I'm afraid of all of the modern world especially in
| technology, I guess if now I would "come back" to all of
| modern and new things: the commercialized world, AI,
| corporations, etc...my head would explode. I mean I can't
| imagine living in such world. I am not sure if everything
| would be alright eith myself in all this everything,This is
| just too much...
| bg24 wrote:
| I am on a max subscription for Claude, and hate the fact that
| OpenAI have not figured out that $20 => $200 is a big jump. Good
| luck to them. In terms of model, just last night, Codex 5.2
| solved a problem for me which other models were going round and
| round. Almost same instructions. That said, I still plan to be on
| $100 Claude (overall value across many tasks, ability to create
| docs, co-work), and may bump up OpenAI subscription to the next
| tier should they decide to introduce one. Not going to $200 even
| with 5.3, unless my company pays for it.
| andix wrote:
| I guess the jump is on purpose. You can buy Codex credits and
| also use codex via the API (manual switching required).
| aerhardt wrote:
| I'm coding about 6-9h per day with Codex CLI on the $20 Plus
| sub, occasionally switching to extra-high reasoning and feeding
| it massive contexts, all tools enabled, sometimes 2-3 terminal
| sessions running in parallel and I've never hit limits... I
| operate on small-ish codebases but even so I try to work in the
| most local scope possible with AGENTS.md at the sub-directory
| levels.
|
| Are you really hitting limits, or are you turned off by the
| fact you _think_ you will?
| bg24 wrote:
| You are correct :-) I am turned off by the fact that I will
| hit the limit if I used more. But you gave me confidence. I
| guess $20 can go a long way. I think only once in the last 3
| months I got rate limited in Codex.
| wiether wrote:
| I use Codex in OpenCode through the API and find the experience
| quite enjoyable.
| bg24 wrote:
| Need to try OpenCode. Thanks.
| satvikpendem wrote:
| You should look into Kilo Pass by Kilo Code
| (https://kilo.ai/features/kilo-pass). It's basically a fixed
| subscription and your credits roll over each month, and you get
| free extra credits too which are used up first before paid
| credits. It's similar to paying for Cursor except the credits
| roll over which is why I'm contemplating moving to it, because
| I don't want to be locked into any one LLM provider the way
| Claude Code or Codex make you become.
| bryanhogan wrote:
| The most important question: Can it do Svelte now?
| speedgoose wrote:
| Today is the best day to rewrite everything in React. You may
| not enjoy React, but AI agents do. And they are the ones
| writing the code.
| davidmurdoch wrote:
| 5.2 was already very good with svelte 5, at least when you have
| the svelte MCP server set up.
| davidmurdoch wrote:
| I've been using 5.2 the way they're describing the new use case
| for 5.3 this whole time.
| tyfon wrote:
| I'm having a hard time parsing the openai website.
|
| Anyone know if it is possible to use this model with opencode
| with the plus subscription?
| karmasimida wrote:
| For those who cared:
|
| GPT-5.3-Codex dominates terminal coding with a roughly 12% lead
| (Terminal-Bench 2.0), while Opus 4.6 retains the edge in general
| computer use by 8% (OSWorld).
|
| Anyone knows the difference between OSWorld vs OSWorld Verified?
| nopinsight wrote:
| From Claude 4.6 Thinking:
|
| OSWorld is the full 369-task benchmark. OSWorld Verified is a
| ~200-task subset where humans have confirmed the eval scripts
| reliably score success/failure -- the full set has some noisy
| grading where correct actions can still get marked wrong.
|
| Scores on Verified tend to run higher, so they're not directly
| comparable.
| bgirard wrote:
| > Using the develop web game skill and preselected, generic
| follow-up prompts like "fix the bug" or "improve the game",
| GPT-5.3-Codex iterated on the games autonomously over millions of
| tokens.
|
| I wish they would share the full conversation, token counts and
| more. I'd like to have a better sense of how they normalize these
| comparisons across version. Is this a 3-prompt 10m token game? a
| 30-prompt 100m token game? Are both models using similar
| prompts/token counts?
|
| I vibe coded a small factorio web clone [1] that got pretty far
| using the models from last summer. I'd love to compare against
| this.
|
| [1] https://factory-gpt.vercel.app/
| veb wrote:
| I just wanted to say that's a pretty cool demo! I hadn't
| realised people were using it for things like this.
| bgirard wrote:
| Thank you. There's a demo save to get the full feel of it
| quickly. There's also a 2D-ASCII and 3D render you can
| hotswap between. The 3D models are generated with Meshy. The
| entire game is 'AI slop'. I intentionally did no code reviews
| to see where that would get me. Some prompts were very
| specific but other prompts were just 'add a research of your
| choice'.
|
| This was built using old versions of Codex, Gemini and
| Claude. I'll probably work on it more soon to try the latest
| models.
| nananana9 wrote:
| I've been listening to the insane 100x productivity gains you all
| are getting with AI and "this new crazy model is a real game
| changer" for a few years now, I think it's about time I asked:
|
| Can you guys point me ton a single useful, majority LLM-written,
| preferably reliable, program that solves a non-trivial problem
| that hasn't been solved before a bunch of times in publicly
| available code?
| eviks wrote:
| Great question, here is the link from the future:
| beernet wrote:
| Can you point me to a human written program an LLM cannot
| write? And no, just answering with a massively large codebase
| does not count because this issue is temporary.
|
| Some people just hate progress.
| falloutx wrote:
| Even a normal website like landonorris.com. Try copying all
| those effects with AI.
|
| Another example: Red Dead Redemption 2
|
| Another one: Roller coaster tycoon
|
| Another one: ShaderToy
| satvikpendem wrote:
| Why do you believe an LLM can't write these, just because
| they're 3D? If the assets are given (just as with a human
| game programmer, who has artists provide them the assets),
| then an LLM can write the code just the same.
| falloutx wrote:
| What? People can easily get assets, thats not a even a
| problem in 2026. Roller coaster tycoon's assets were done
| by the programmer himself. If its so easy why haven't we
| seen actually complex pieces of software done in couple
| of weeks by LLM users?
|
| Also try building any complex effects by prompting LLMs,
| you wont get any far, this is why all of the LLM coded
| websites look stupidly bland.
| satvikpendem wrote:
| Not sure what you're confused about, I never said assets
| were hard to get, I just said that the LLM needs to be
| provided a folder of the assets for it to make use of
| them, it's not going to create them from scratch (at
| least not without great difficulty, because LLMs are
| capable of using and coding Three.js for example). I
| don't know the answer to your first question because I
| don't hang around in the 3D or game dev fields, I'm sure
| there are examples of vibe coded games however.
|
| As to your second question, it is about prompting them
| correctly, for example [0]. Now I don't know about you
| but some of those sites especially after using the
| frontend skill look pretty good to me. If those look
| bland to you then I'm not really sure what you're
| expecting, keeping in mind that the example you showed
| with the graphics are not regular sites but more design
| oriented, and even still nothing stops LLMs from
| producing such sites.
|
| [0] https://youtu.be/f2FnYRP5kC4
| falloutx wrote:
| you have shown me 0 examples, I showed actual examples to
| the given question. Your answers have just been "AI can
| also do this" but gave no actual proof.
| satvikpendem wrote:
| The examples are in the video I linked, as I said, if you
| don't bother to watch it then I'm not sure what to tell
| you. As I said for games I don't know and won't presume
| to search up some random vibe coded game if I don't have
| personal experience with how LLMs handle games, but for
| web development, the sites I've made and seen made look
| pretty good.
| avaer wrote:
| I wish I could agree with you, but as a game dev, shader
| author, and occasional asm hacker, I still think AIs have
| demonstrated being perfectly capable of copying "those
| effects". It's been trained on them, of course.
|
| You're not gonna one-shot RD2, but neither will a human.
| You can one-shot particles and shader passes though.
| falloutx wrote:
| I didnt say one shot it, coding agents have been out for
| more than couple years and yet we cant point to single
| Good piece of software built by it.
| suddenlybananas wrote:
| And some people clearly hate humans.
| svantana wrote:
| Pretty much any software that people pay for? If LLMs could
| clone an app, why would anyone still pay good money for the
| original?
| HAL3000 wrote:
| > Can you point me to a human written program an LLM cannot
| write?
|
| Sure:
|
| "The resulting compiler has nearly reached the limits of
| Opus's abilities. I tried (hard!) to fix several of the above
| limitations but wasn't fully successful. New features and
| bugfixes frequently broke existing functionality.
|
| As one particularly challenging example, Opus was unable to
| implement a 16-bit x86 code generator needed to boot into
| 16-bit real mode. While the compiler can output correct
| 16-bit x86 via the 66/67 opcode prefixes, the resulting
| compiled output is over 60kb, far exceeding the 32k code
| limit enforced by Linux. Instead, Claude simply cheats here
| and calls out to GCC for this phase (This is only the case
| for x86. For ARM or RISC-V, Claude's compiler can compile
| completely by itself.)"[1]
|
| 1. https://www.anthropic.com/engineering/building-c-compiler
| xandrius wrote:
| Why even come to this site if you're so anti-innovation?
|
| Today with LLMs you can literally spend 5 minutes defining what
| you want to get, press send, go grab a coffee and come back to
| a working POC of something, in literally any programming
| language.
|
| This is literally stuff of wonders and magic that redefines how
| we interface with computers and code. And the only thing you
| can think of is to ask if it can do something completely novel
| (that it's so hard to even quantity for humans that we don't
| have software patents mainly for that reason).
|
| And the same model can also answer you if you ask it about
| maths, making you an itinerary or a recipe for lasagnas. C'mon
| now.
| svantana wrote:
| There are different kinds of innovation.
|
| I want AI that cures cancer and solves climate change.
| Instead we got AI that lets you plagiarize GPL code, does
| your homework for you, and roleplay your antisocial horny
| waifu fantasies.
| legulere wrote:
| I don't think that the user you are responding to is anti-
| innovation, but rather points out that the usefulness of AI
| is oversold.
|
| I'm using Copilot for Visual Studio at work. It is useful for
| me to speed some typing up using the auto-complete. On the
| other hand in agentic mode it fails to follow simple basic
| orders, and needs hand-holding to run. This might not be the
| most bleeding-edge setup, but the discrepancy between how
| it's sold and how much it actually helps for me is very real.
| ifwinterco wrote:
| I think copilot is widely considered to be fairly rubbish,
| your description of agentic coding was also my experience
| prior to ~Q3 2025, but things have shifted meaningfully
| since then
| pkoiralap wrote:
| In the 1930s, when electronic calculators were first
| introduced, there was a widespread belief that accounting as a
| career was finished. Instead, the opposite became true.
| Accounting as a profession grew, becoming far more
| analytical/strategic than it had been previously.
|
| You are correct that these models primarily address problems
| that have already been solved. However, that has always been
| the case for the majority of technical challenges. Before LLMs,
| we would often spend days searching Stack Overflow to find and
| adapt the right solution.
|
| Another way to look at this is through the lens of problem
| decomposition as well. If a complex problem is a collection of
| sub-problems, receiving immediate solutions for those
| components accelerates the path to the final result.
|
| For example, I was recently struggling with a UI feature where
| I wanted cards to follow a fan-like arc. I couldn't quite get
| the implementation right until I gave it to Gemini. It didn't
| solve the entire problem for me, but it suggested an approach
| involving polar coordinates and sine/cosine values. I was able
| to take that foundational logic turn it into a feature I
| wanted.
|
| Was it a 100x productivity gain? No. But it was easily a 2x
| gain, because it replaced hours of searching and waiting for a
| mental breakthrough with immediate direction.
|
| There was also a relevant thread on Hacker News recently
| regarding "vibe coding":
|
| https://news.ycombinator.com/item?id=45205232
|
| The developer created a unique game using scroll behavior as
| the primary input. While the technical aspects of scroll events
| are certainly "solved" problems, the creative application was
| novel.
| suddenlybananas wrote:
| The story you're describing doesn't seem much better than one
| could get from googling around and going on stackoverflow
| strokirk wrote:
| It doesn't have to be, really. Even if it could replace 30%
| of documentation and SO scrounging, that's pretty valuable.
| Especially since you can offload that and go take a coffee.
| rohit89 wrote:
| > that hasn't been solved before a bunch of times in publicly
| available code?
|
| And this matters because? Most devs are not working on novel
| never before seen problems.
| kevstev wrote:
| Heh, I agree. There is a vast ocean of dev work that is just
| "upgrade criticalLib to v2.0" or adding support for a new
| field from the FE through to the BE.
|
| I can name a few times where I worked on something that you
| could consider groundbreaking (for some values of
| groundbreaking), and even that was usually more the
| combination of small pieces of work or existing ideas.
|
| As maybe a more poignant example- I used to do a lot of on-
| campus recruiting when I worked in HFT, and I think I
| disappointed a lot of people when I told them my day to day
| was pretty mundane and consisted of banging out Jiras,
| usually to support new exchanges, and/or securities we hadn't
| traded previously. 3% excitement, 97% unit tests and covering
| corner cases.
| revahage wrote:
| Well, it took opus 4.5 five messages to solve a trivial git
| problem for me. It hallucinated nonexistent flags three times.
| Hallucinating nonexistent flags is certainly a novel solution
| to my git ineptness.
|
| Not to be outdone, chatgpt 5.2 thinking high only needed about
| 8 iterations to get a mostly-working ffmpeg conversion script
| for bash. It took another 5 messages to translate it to run in
| windows, on powershell (models escaping newlines on windows
| properly will be pretty nuch AGI, as far as I'm concerned).
| Def_Os wrote:
| Yeah, I would LOVE to see attempts at significant video games
| that are then open-sourced for communities to work on. E.g.
| OpenGTA or OpenFIFA/OpenNHL.
| llmslave wrote:
| baffled that people are still suspicious of ai coding models
| PieUser wrote:
| How'd they both release at the same time? Insiders?
| Rperry2174 wrote:
| Whats interesting to me is that these gpt-5.3 and opus-4.6 are
| diverging philosophically and really in the same way that actual
| engineers and orgs have diverged philosophically
|
| With Codex (5.3), the framing is an interactive collaborator: you
| steer it mid-execution, stay in the loop, course-correct as it
| works.
|
| With Opus 4.6, the emphasis is the opposite: a more autonomous,
| agentic, thoughtful system that plans deeply, runs longer, and
| asks less of the human.
|
| that feels like a reflection of a real split in how people think
| llm-based coding should work...
|
| some want tight human-in-the-loop control and others want to
| delegate whole chunks of work and review the result
|
| Interested to see if we eventually see models optimize for those
| two philosophies and 3rd, 4th, 5th philosophies that will emerge
| in the coming years.
|
| Maybe it will be less about benchmarks and more about different
| ideas of what working-with-ai means
| karmasimida wrote:
| > With Codex (5.3), the framing is an interactive collaborator:
| you steer it mid-execution, stay in the loop, course-correct as
| it works.
|
| > With Opus 4.6, the emphasis is the opposite: a more
| autonomous, agentic, thoughtful system that plans deeply, runs
| longer, and asks less of the human.
|
| Ain't the UX is the exact opposite? Codex thinks much longer
| before gives you back the answer.
| xd1936 wrote:
| I've also had the exact opposite experience with tone. Claude
| Code wants to build with me, and Codex wants to go off on its
| own for a while before returning with opinions.
| mrkstu wrote:
| Its likely that both are steering towards the middle from
| their current relative extremes and converging to nearly
| the same place.
| gervwyk wrote:
| also my experience in using these two models. they are
| trying to recover from oversteer perhaps.
| WilcoKruijer wrote:
| Yes, you're right for 4.5 and 5.2. Hence they're focusing on
| improving the opposite thing and thus are actually
| converging.
| bt1a wrote:
| This is most likely an inference serving problem in terms of
| capacity and latency given that Opus X and the latest GPT
| models available in the API have always responded quickly and
| slowly, respectively
| cwyers wrote:
| Codex now lets you tell the LLM tgings in the middle of its
| thinking without interrupting it, so you can read the
| thinking traces and tell it to change course if it's going
| off track.
| utilize1808 wrote:
| I think it's the opposite. Especially considering Codex started
| out as a web app that offers very little interactivity: you are
| supposed to drop a request and let it run automatously in a
| containerized environment; you can then follow up on it via
| chat --- no interactive code editing.
| Rperry2174 wrote:
| Fair I agree that was true of early codex and my perception
| too.. but today there are two announcements that came out and
| thats what im referring to.
|
| specifically, the GPT-5.3 post explicitly leans into
| "interactive collaborator" langauge and steering mid
| execution
|
| OpenAI post: "Much like a colleague, you can steer and
| interact with GPT-5.3-Codex while it's working, without
| losing context."
|
| OpenAI post: "Instead of waiting for a final output, you can
| interact in real time--ask questions, discuss approaches, and
| steer toward the solution"
|
| Claude post: "Claude Opus 4.6 is designed for longer-running,
| agentic work -- planning complex tasks more carefully and
| executing them with less back-and-forth from the user."
| d--b wrote:
| I am definitely using Opus as an interactive collaborator that
| I steer mid-execution, stay in the loop and course correct as
| it works.
|
| I mean Opus asks a lot if he should run things, and each time
| you can tell it to change. And if that's not enough you can
| always press esc to interrupt.
| mcintyre1994 wrote:
| This kind of sounds like both of them stepping into the other's
| turf, to simplify a bit.
|
| I haven't used Codex but use Claude Code, and the way people
| (before today) described Codex to me was like how you're
| describing Opus 4.6
|
| So it sounds like they're converging toward "both these
| approaches are useful at different times" potentially? And
| neither want people who prefer one way of working to be locked
| to the other's model.
| cchance wrote:
| Just because you can inject steering doesn't mean they stered
| away from long running...
|
| Theres hundreds of people who upload Codex 5.2 running for
| hours unattended and coming back with full commits
| giancarlostoro wrote:
| > With Opus 4.6, the emphasis is the opposite: a more
| autonomous, agentic, thoughtful system that plans deeply, runs
| longer, and asks less of the human.
|
| This feels wrong, I can't comment on Codex, but Claude will
| prompt you and ask you before changing files, even when I run
| it in dangerous mode on Zed, I can still review all the diffs
| and undo them, or you know, tell it what to change. If you're
| worried about it making too many decisions, you can pre-prompt
| Claude Code (via .claude/instructions.md) and instruct it to
| always ask follow up questions regarding architectural
| decisions.
|
| Sometimes I go out of my way to tell Claude DO NOT ASK ME FOR
| FOLLOW UPS JUST DO THE THING.
| Rperry2174 wrote:
| yeah I'm mostly just talking about how they're framing it:
| "Claude Opus 4.6 is designed for longer-running, agentic work
| -- planning complex tasks more carefully and executing them
| with less back-and-forth from the user"
|
| I guess its also quite interesting that how they are framing
| these projects are opposite from how people currently
| perceive them and I guess that may be a conscious choice...
| giancarlostoro wrote:
| I get what you mean now, I like that to be fair, sometimes
| I want Claude to tell me some architectural options, so I
| ask it so I can think about what my options are, sometimes
| I rethink my problem if I like Claudes conclusion.
| ghosty141 wrote:
| I'm personally 100% convinced (assuming prices stay reasonable)
| that the Codex approach is here to stay.
|
| Having a human in the loop eliminates all the problems that
| LLMs have and continously reviewing small'ish chunks of code
| works really well from my experience.
|
| It saves so much time having Codex do all the plumbing so you
| can focus on the actual "core" part of a feature.
|
| LLMs still (and I doubt that changes) can't think and
| generalize. If I tell Codex to implement 3 features he won't
| stop and find a general solution that unifies them unless
| explicitly told to. This makes it kinda pointless for the "full
| autonomy" approach since effecitly code quality and
| abstractions completely go down the drain over time. That's
| fine if it's just prototyping or "throwaway" scripts but for
| bigger codebases where longevity matters it's a dealbreaker.
| _zoltan_ wrote:
| I'm personally 100% convinced of the opposite, that it's a
| waste of time to steer them. we know now that agentic loops
| can converge given the proper framing and self-reflectiveness
| tools.
| sealeck wrote:
| Converge towards what though... I think the level of
| testing/verification you need to have an LLM output a non-
| trivial feature (e.g. Paxos/anything with concurrency,
| business logic that isn't just "fetch value from
| spreadsheet, add to another number and save to the
| database") is pretty high.
| techbro_1a wrote:
| > With Codex (5.3), the framing is an interactive collaborator:
| you steer it mid-execution, stay in the loop, course-correct as
| it works.
|
| This is true, but I find that Codex thinks more than Opus.
| That's why 5.2 Codex was more reliable than Opus 4.5
| rozumbrada wrote:
| I read this exact comment with I would say completely the same
| words several times in X and I would bet my money it's LLM
| generated by someone who has not even tried both the tools.
| This AI slop even in the site like this without direct
| monetisation implications from fake engagement is making me
| sick...
| blurbleblurble wrote:
| Funny cause the situation was totally flipped last iteration.
| bob1029 wrote:
| I think there is another philosophy where the agent is domain
| specific. Not that we have to invent an entirely new universe
| for every product or business, but that there is a small amount
| of semi-customization involved to achieve an ideal agent.
|
| I would much rather work with things like the Chat Completion
| API than any frameworks that compose over it. I want total
| control over how tool calling and error handling works. I've
| got concerns specific to my business/product/customer that
| couldn't possibly have been considered as part of these
| frameworks.
|
| Whether or not a human needs to be tightly looped in could vary
| wildly depending on the specific part of the business you are
| dealing with. Having a purpose-built agent that understands
| where additional verification needs to occur (and not occur)
| can give you the best of both worlds.
| hbarka wrote:
| How can they be diverging, LLMs are built on similar
| foundations aka the Transformer architecture. Do you mean the
| training method (RLHF) is diverging?
| iranintoavan wrote:
| I'm not OP but I suspect they are meaning the products /
| tooling / company direction, not necessarily the underlying
| LLM architecture.
| pyrolistical wrote:
| Boing vs airbus philosophy
| jhancock wrote:
| Good breakdown.
|
| I usually want the codex approach for code/product "shaping"
| iteratively with the ai.
|
| Once things are shaped and common "scaling patterns" are well
| established, then for things like adding a front end (which is
| constantly changing, more views) then letting the autonomous
| approach run wild can *sometimes* be useful.
|
| I have found that codex is better at remembering when I ask to
| not get carried away...whereas claude requires constantly
| reminders.
| gallerdude wrote:
| Both Opus 4.6 and GPT-5.3 one shot a Gameboy emulator for me.
| Guess I need a better benchmark.
| paxys wrote:
| As coding agents get "good enough" the next differentiator will
| be which one can complete a task in fewer tokens.
| tgtweak wrote:
| Or quicker, or more comprehensively for the same price.
| nlh wrote:
| Or the same number of tokens in less time. Kinda feels like
| the CPU / modem wars of the 90s all over again - I remember
| those differences you felt going from a 386 -> 486 or from a
| 2400 -> 9600 baud modem.
|
| We're in the 2400 baud era for coding agents and I for one
| look forward to the 56k era around the corner ;)
| well_ackshually wrote:
| There's hundreds of gameboy emulators available on Github
| they've been trained on. It's quite literally the simplest
| piece of emulation you could do. The fact that they couldn't do
| it before is an indictment of how shit they were, but a gameboy
| emulator should be a weekend project for anyone even ever so
| slightly qualified. Your benchmark was awful to begin with.
| gf000 wrote:
| Is such an emulator not part of their training data sets?
| nubg wrote:
| lmao so cringe that they delay releasing the model until
| anthropic does
| dllrr wrote:
| Using opus 4.6 in claude code right now. It's taking about 5x
| longer to think things through, if not more.
| andyferris wrote:
| The notes explicitly call out you may want to dial the effort
| setting back to medium to reduce latency/tokens (high being
| default, apparently there is a max setting too).
| jpau wrote:
| Interesting that this was released without a prior GPT-5.3
| release. I wonder if that means we won't see a GPT-5.3?
| koolala wrote:
| I want to recompile a Rust project to be f32 instead of f64.
|
| Am I better off buying 1 month of Codex, Claude, or Antigravity?
|
| I want to have the agent continuesly recompile and fix compile
| errors on loop until all the bugs from switching to f32 are gone.
| EmilStenstrom wrote:
| Doesn't matter which one. All of them can do things like this
| now, given a good enough feedback loop. Which your problem has.
| argsnd wrote:
| All of them can do it but Codex has the least frustrating usage
| limits.
| koolala wrote:
| When using it in VSCode? The browser system running its own
| container seems like it would be the most demanding on their
| resources. The stand-alone client is Mac-only but I don't
| know if it makes a difference.
|
| My goal is to do it within the usage I get from a $20 monthly
| plan.
| vatsachak wrote:
| Literally just find and replace
| koolala wrote:
| find and replace is step 1 that generates all the compile
| errors I want it to loop through
|
| I'm wanting to do it on an entire programming language made
| in rust: https://github.com/uiua-lang/uiua
|
| Because there are no float32 array languages in existence
| today
| SunshineTheCat wrote:
| I've always been fascinated to see significantly more people
| talking about using Claude than I see people talking about Codex.
|
| I know that's anecdotal, but it just seems Claude is often the
| default.
|
| I'm sure there are key differences in how they handle coding
| tasks and maybe Claude is even a little better in some areas.
|
| However, the note I see the most from Claude users is running out
| of usage.
|
| Coding differences aside, this would be the biggest factor for me
| using one over the other. After several months on Codex's $20/mo.
| plan (and some pretty significant usage days), I have only come
| close to my usage limit once (never fully exceeded it).
|
| That (at least to me) seems to be a much bigger deal than coding
| nuances.
| fHr wrote:
| Codex is great and I hit the usage once doing multiagent full 5
| hour absolute degen session for the nornal workflow alongside
| never hit it and now x2 useage even and now with the planmode
| switch back and forth absolute great.
| AstroBen wrote:
| I'm with you. Codex's plans seems to be a much more generous
| offering than Claude
|
| I just.. can't tell a different in quality between them.. so I
| go for the cheapest
| superfrank wrote:
| I only switched to using the terminal based agents in the last
| week. Prior to this I was pretty much only using it through
| Cursor and GH Copilot. The Anthropic models when used through
| GH Copilot were far superior to the codex ones and I didn't
| really get the hype of Codex. Using them through the CLI
| though, Codex is much better, IMO.
|
| My guess is that it's potentially that and just momentum from
| developers who started using CC when it was far superior to
| Codex has allowed it to become so much more popular.
| Potentially, it's might be that, as it's more autonomous, it's
| better for true vibe-coding and it's more popular with the
| Twitter/LinkedIn wantrepreneur crew which meant it gets a lot
| of publicity which increases adoption quicker.
| timpera wrote:
| In my experience, OpenAI gives you unreasonable amounts of
| compute for EUR20/month. I am subscribed to both and Claude's
| limits are so tiny compared to ChatGPT's that it often feels
| like a rip-off.
|
| Claude also doesn't let you use a worse model after you reach
| your usage limits, which is a bit hard to swallow when you're
| paying for the service.
| mrandish wrote:
| > the note I see the most from Claude users is running out of
| usage.
|
| I suspect that tells us less about model capability/efficiency
| and more about each company's current need to paint a specific
| picture for investors re: revenue, operating costs, capital
| requirements, cash on hand, growth rate, retention, margins
| etc. And those needs can change at any moment.
|
| Use whatever works best for your particular needs today, but
| expect the relative performance and value between leaders to
| shift frequently.
| petetnt wrote:
| Whoa, I think GPT-5.2-Codex was a disappointment, but
| GPT-5.3-Codex is definitely the future!
| drcongo wrote:
| Does it insert adverts in your code?
| mrcwinn wrote:
| According to Sam Altman, Anthropic is for "rich people." Judging
| by his $4 million man-baby Koeniggsegg, he must be a huge Claude
| Code user!
| RivieraKid wrote:
| Do software engineers here feel threatened by this? I certainly
| am. I'm surprised that this topic is almost entirely missing in
| these threads.
| ReptileMan wrote:
| Jevons paradox hints that the situation is not as bleak as it
| sounds.
| worldsavior wrote:
| No. AI does not work well enough, you still need a person to
| look on it and CODE. It probably never will, until AGI which
| probably also in my opinion will never come.
| dude250711 wrote:
| It's a super-special AI tier that can replace developers and
| other grunts, yet somehow cannot replace managers and
| C-suite.
|
| It can only replace whoever is not writing a fat cheque to
| it.
| OsrsNeedsf2P wrote:
| I would feel threatened if I didn't invest in learning how to
| best use AI
| vatsachak wrote:
| AI is mostly garbage at creating useful abstractions. I'd feel
| threatened if I was a competitive programmer or IMO kid
| llmslave wrote:
| theres alot of denial, and people that havent taken a serious
| look at the ai models
| virtualzx wrote:
| is so fun that the two releases used almost completely non-
| overlapping benchmarks!
| vatsachak wrote:
| AI designed websites are so easy to spot that I need to actively
| design my UI so that it doesn't look AI
| textlapse wrote:
| I would love to see a nutritional facts label on how many prompts
| / % of code / ratio of human involvement needed to use the models
| to develop their latest models for the various parts of their
| systems.
| jiggawatts wrote:
| I think this announcement says a lot about OpenAI and their
| relationship to partners like Microsoft and NVIDIA, not to
| mention the attitude of their leadership team.
|
| On Microsoft Foundry I can see the new Codex 4.6 model _right
| now_ , but GPT-5.3 is nowhere to be seen.
|
| I have a pre-paid account directly with OpenAI that has credits,
| but if I use that key with the Codex CLI, it can't access 5.3
| either.
|
| The press release very prominently includes this quote:
| "GPT-5.3-Codex was co-designed for, trained with, and served on
| NVIDIA GB200 NVL72 systems. We are grateful to NVIDIA for their
| partnership."
|
| Sounds like OpenAI's ties with their vendors are fraying while at
| the same time they're struggling to execute on the basics like
| "make our own models available to our own coding agents", let
| alone via third-party portals like Microsoft Foundry.
| tombert wrote:
| Actually kind of excited for this. I've been using 5.2 for awhile
| now, and it's already pretty impressive if you set the context
| window to "high".
|
| Something I have been experimenting with is AI-assisted proofs.
| Right now I've been playing with TLAPS to help write some more
| comprehensive correctness proofs for a thing I've been building,
| and 5.2 didn't seem _quite_ up to it; I was able to figure out
| proofs on my own a bit better than it was, even when I would tell
| it to keep trying until it got it right.
|
| I'm excited to see if 5.3 fairs a bit better; if I can get
| mechanized proofs working, then Fields Medal here I come!
| sidgarimella wrote:
| Many are saying codex is more interactive but ironically I think
| that very interactivity/determinism works best when using codex
| remotely as a cloud agent and in highly async cases. Conversely I
| find opus great locally, where I can ram messages into it to try
| to lever its autonomy best (and interrupt/clean up)
___________________________________________________________________
(page generated 2026-02-05 23:00 UTC)