We have partnered with the Department of War to deploy our models on their classified network, ensuring all use adheres to strict AI safety principles, including human oversight of force and a ban on mass surveillance.

We love working with NVIDIA, and they make the best AI chips in the world. We hope to be a gigantic customer for a very long time. I don't get where all this insanity is coming from.

It has been amazing to watch the progress of the Codex team; they are beasts. The product/model is already so good and will get much better; I believe they will create the best and most important product in the space, and enable so much downstream work.

Very pleased to be working with Amazon to bring a lot more NVIDIA chips online for OpenAI to keep scaling!

Sora update #1: blog.samaltman.com/sora-update-...

ChatGPT's GPT-5 update offers Auto, Fast, and Thinking modes, with a 3,000 messages/week limit for Thinking and a 196k token context. GPT-4o returns for paid users, with more models accessible. A warmer GPT-5 personality and user customization are planned. #ChatGPT

Most users should like GPT-5 better soon; the change is rolling out over the next days. The real solution here remains letting users customize #ChatGPT's style much more. We are working that!