Had such a great time at @vercel Ship SF yesterday! 🙌
Thanks to @LukePS for co-presenting, the Vercel team for hosting, and everyone who joined “Embed Gen UI Agents in Your App with Codex and Vercel.”
We explored agents that can dynamically build interactive UI/UX on the fly based on user inputs. More info in the thread below 👇
Great day at the Agents booth at OpenAI DevDay! Loved meeting everyone and hearing all that you are building. Lots of excitement around Agents API and ultrafast Astra in the API.
Then Agents API is an OpenAI-hosted Codex harness accessible via API, making it easier than ever to build agents and embed the power of Codex into your applications.
Ultrafast is a new API service tier up to 8x faster than standard tier, unlocking new UI/UX experiences that were not possible before (near real-time GenUI etc.). Will share a demo on this soon.
More details below 👇
Please welcome GPT-6 Sol and GPT-6 Luna to the GPT-6 universe.
GPT-6 Sol and Luna build on the advances behind GPT-6 Astra, bringing much of its strengths into faster and more affordable models to support work at scale.
We’ve also made caching and inference more efficient, and we’re passing the savings directly to you: 50% lower API prices for Sol and Luna compared with GPT‑5.6 promotional pricing.
I’ve been experimenting with GPT-6 Astra in Codex for video creation/editing, and it’s been working super well! For this video, I asked it to create an AI avatar of me with a clone of my voice. More details below 👇
A few tips if you want to try something similar:
• Define the audience, target length and where you’ll post upfront
• Share a reference video and specify what to match: pacing, framing, and/or captions
• Approve the script, then the voice narration, before generating the avatar to avoid costly voice/avatar changes later
• Ask for a rough cut first; get the pacing right before adding graphics or captions
• Give precise edit notes: “remove the pause at 0:08, but keep the full sentence”
• Keep the editable project and style settings so future videos need less setup
• Turn what worked well into a Skill for Codex to use again next time; you can ask Codex to do this for you!
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.
Launch blog : https://t.co/SAdR7fw8zS
If you're building with the API, the model/prompting guide is a great resource to get started:
https://t.co/BLtfJZGhBp
Astra is SoTA in computer use, browsing, software engineering and more, scoring 99.9% on ARC-AGI-3!
To the stars! ✨
GPT-6 Astra went live for all today in the API and is rolling out across ChatGPT/Codex! Details below👇
I'm super inspired by all the cool demos people are sharing!
If you are working with Astra, share a demo or DM me! Would love to learn more about what you're building.