Every few months, we take another leap that would have sounded ridiculous a year ago.
What one person can build now is truly just bottlenecked by their imagination.
GPT-6-Astra is here and I am very curious to see what you do with it and hear feedback on how we can improve it!
Sometimes all you can do is keep your heads down and ship valuable things. So so proud of the team(s) that worked on GPT-5.4. The wide reception is blowing us all away and more than makes up for all the sleepless nights.
DMs open for any feedback!
It's potentially the most competitive last 30 days in AI Coding history.
GPT-5.2 represents the biggest leap for GPT models in agentic coding since GPT-5 and is a SOTA coding model in its price range. The version bump undersells the jump in intelligence. We’re excited to make it the default across Windsurf and several core Devin workloads.
Couldn't be more proud to the teams involved (s/o to power-users and frontiers). I am constantly amazed and inspired by working with @ericmitchellai@yanndubs the absolute lunatics that just refuse to say no (and refuse to sleep).
GPT-5.2 is out! Absolutely crazy the amount of progress we are making literally week over week.
Evals only tell part of the story and don't be fooled by the 0.1 increment in the model naming.
It both extends the ceiling of AI capability and improves on day-to-day interactivity.
GPT-5.1 is out! It's a nice upgrade.
I particularly like the improvements in instruction following, and the adaptive thinking.
The intelligence and style improvements are good too.
That said, the job is (obviously) not done. There lies a lot of (exciting) work ahead of us and we will keep striving to make the model better and better, and have it provide more and more utility to everyone.
GPT-5 is finally out! Amongst its many improvements, I am personally very excited for everyone to try its enhanced coding abilities — both agenetic and front-end.
And give it your hardest tasks in your favourite IDE or vibe code feature rich and tasteful webapps.
I had the opportunity to work with some of the smartest, most passionate group of cracked researchers and engineers. Folks who truly care about the craft, and I am both incredibly proud and honoured to have been part of this GPT-5 development journey.
1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).
GPT-4.1 (and -mini and -nano) are now available in the API!
these models are great at coding, instruction following, and long context (1 million tokens).
benchmarks are strong, but we focused on real-world utility, and developers seem very happy.
GPT-4.1 family is API-only.
i'm also incredibly proud of the team to get to this milestone. it really takes a village to release something like this. in particular, really proud of my team on post training: