Incredible times!
We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics.
The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra.
The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.
Florent Bekerman retweeted
I would like to clarify a few things:
1) The screenshot is my reaching out to Levent to coordinate our releases. I hope it’s clear from the message that we came in with the best possible intentions.
2) I never ever asked for Levent to be removed from authorship of his own work (as indicated by my text). I was surprised to learn during the call with Tristan that they had only solved Euler and not Navier-Stokes; after learning this we brainstormed possible paths forward. One option we discussed was that Tristan could be the lead author on a rewrite of OpenAI’s Navier-Stokes proof. It is in that context that I said “it would be simpler if Levent was not an Anthropic employee” because I felt it would be inappropriate for an Anthropic employee to author OpenAI’s work. Importantly it was admitted that internal Anthropic models had been used in their proof of Euler blowup; I therefore felt I could not consider Levent to be an independent academic. Another option I wanted to propose (but got cut short) is to offer access to our internal model so that they could try to finish their proof and bridge the gap between Euler and NS. Again I did not know how to navigate giving access to internal OpenAI IP to an Anthropic employee.
3) To reiterate it plainly: as my text clearly indicates, and as I said during our call, OpenAI's intention was to do everything possible to celebrate their mathematical achievements and the heroic efforts that they made on Euler. In the call I was immediately met with a litany of slander, including direct threats that if we were to announce Navier-Stokes he would immediately go to the press with a barrage of unfounded accusations. I refuted all these accusations but he replied “there is nothing you can do, I simply do not trust you”. I was confused why one would turn an incredible source for celebration (of their achievements!) into such bickering, which is when I said that I did not understand why one would risk their career [over unfounded accusations]. Genuinely, at that moment, I was trying to care for him and do a last ditch attempt to get a chance to give them all the credits that they deserve. I deeply apologize for this extremely poor choice of words, it is the opposite of what I was trying to convey. (I should say that I retracted them on the spot by the way.)
4) Overall, on a personal level, it was incredibly difficult to have these conversations. Levent refused to attend any of the meetings despite my repeated asking. As Sholto Douglas said, there will need to be coordination between Anthropic and OpenAI in the future; I felt I was doing a proxy negotiation with Anthropic while the Anthropic employee refused to directly participate.
Very important essay from Jakub (our Chief Scientist at OpenAI)
« The intelligence produced by scaling deep learning is not directly comparable to human intelligence. To become very relevant in the real world - very useful or very dangerous - the AI does not need to match or exceed all human capabilities; it just needs to surpass enough of them. And as it continues to surpass humans on more and more axes, it is becoming increasingly difficult to understand exactly how capable it is. »
I wrote about the state of AI, why I’m concerned about the next few years, and the choices we need to make to keep the future in humanity’s hands.
An Alien Mind: openai.com/index/an-alien-mi…
Florent Bekerman retweeted
Improvements for Astra came from more general techniques in development long long before the Hugging Face incident. ExploitGym Honeypot was very recently added as an eval following that incident and is out of distribution for our RL runs. There are also clear improvements across a wider range of behaviors in deployment simulations, deception evals, and realistic computer-use tasks, though these still reveal failures and substantial room for improvement.
We should have definitely done a better job explaining where we think the alignment improvements came from in the system card and that was a miss. Measuring alignment generalization is a core part of our research program - we do not benchmark-maxx alignment evals. This would be horrendously stupid and I hope no lab is doing this.
That being said, metagaming and eval awareness are real challenges for any effort to measure alignment, including ours, and understanding their effects is an active focus. We're spending a lot of effort to improve our eval techniques and study worst-case alignment and metagaming behaviors in our models.
GPT-6 Astra has launched! 6 Astra is a truly incredible model - much more capable, aligned and efficient.
Very excited for everyone to try it out!
Today we released Gpt-5.5!
It’s a considerably more capable and more efficient model. Using this model as a daily driver in Codex has made the pace of progress feel very real.
What I find the most exciting about working at @OpenAI is not only contributing to this progress, but making it accessible to all.
Very proud of the work of teammates and many others to achieve this.
openai.com/index/introducing…
Florent Bekerman retweeted
OK, I think GPT 5.2 Pro is actually a step change in usefulness for my applications (algebraic geometry/number theory research).
Very excited to announce GPT-5.2!
It was an amazing team effort to release this model. What I’m genuinely thrilled about are the improvements at economically valuable tasks - the model has performed at or above the level of human experts 70.9% of the time on a wide range of real world tasks. This reflects what we care the most about and what we’ll keep improving on - providing utility to our users.
And the model does quite well on benchmarks too :)
GPT-5.2 is now rolling out to everyone.
openai.com/index/introducing…
Great take
Social media tends to frame AI debate into two caricatures:
(A) Skeptics who think LLMs are doomed and AI is a bunch of hype.
(B) Fanatics who think we have all the ingredients and superintelligence is imminent.
But if you read what leading researchers actually say (beyond the headlines), there’s a surprising amount of convergence:
1) The current paradigm is likely sufficient for massive economic and societal impact, even without further research breakthroughs.
2) More research breakthroughs are probably needed to achieve AGI/ASI. (Continual learning and sample efficiency are two examples that researchers commonly point to.)
3) We probably figure them out and get there within 20 years. @demishassabis said maybe in 5-10 years. @fchollet recently said about 5 years. @sama said ASI is possible in a few thousand days. @ylecun said about 10 years. @ilyasut said 5-20 years. @DarioAmodei is the most bullish, saying it's possible in 2 years though he also said it might take longer.
None of them are saying ASI is a fantasy, or that it's probably 100+ years away.
A lot of the disagreement is in what those breakthroughs will be and how quickly they will come. But all things considered, people in the field agree on a lot more than they disagree on.
Florent Bekerman retweeted
I crossed an interesting threshold yesterday, which I think many other mathematicians have been crossing recently as well. In the middle of trying to prove a result, I identified a statement that looked true and that would, if true, be useful to me. 1/3
GPT-5 is good at math - despite not using some of the general techniques behind our IMO gold medal model. I’m excited about the future of our models!
I think people are still under indexing on how impressive Nano is. 54% at SWE-Bench (amongst other things) for a model this size is extremely impressive. Congrats @suvansh !
GPT-5 Nano is the cheapest reasoning model offered by any lab – $0.05/$0.40.
It’s incredibly intelligence dense, with over 54% SWE-Bench and nearly 10% FrontierMath on "high" reasoning. No model of any size had reached either mark as recently as January – it beats o1, DeepSeek R1, and Claude 3.6 Sonnet on both evals.
Nano is optimized for function-calling and instruction-following, targeting low-latency, low-cost agentic API use cases. We hope this model (along with GPT-OSS) unlocks new use cases, making reasoning ubiquitous.
And as excited as I am for these small models, I’m more excited for the next ones – we're learning as we go, and there's still low-hanging fruit left.
Florent Bekerman retweeted
gpt-5 is here today! so excited to see what you'll do with it. we've focused a ton on coding in particular, and i think you'll really enjoy using it for the hardest problems and for building web apps end to end. what you can get out of a single prompt is really remarkable
Florent Bekerman retweeted
gpt-5 is now the default in cursor (replacing claude)
"it's the smartest coding model we've tried"
— @mntruell, cursor ceo
Today we are releasing GPT-5! We had a real focus on making the model more factual, reliable and provide real utility to users.
Very proud by the work done by the team here. We think it’s a good model and are excited for everyone to try it. Please give feedback!
The model is meaningfully better at developer use cases, including agentic coding, frontend, function calling and generic tool use. It also happens to do fairly well on benchmarks - check the blog post below :)
openai.com/index/introducing…
