LLMs & Chatbots
Intellectual freedom by design
ChatGPT is designed to be useful, trustworthy, and adaptable—so you can make ...
Expanding on what we missed with sycophancy
A deeper dive on our findings, what went wrong, and future changes we’re making.
From hard refusals to safe-completions: toward outpu...
Discover how OpenAI's new safe-completions approach in GPT-5 improves both sa...
Introducing gpt-oss
We’re releasing gpt-oss-120b and gpt-oss-20b—two state-of-the-art open-weight...
Open Weights and AI for All
AI’s next frontier isn’t just about capability—it’s about who gets to use it....
Figma uses AI to transform digital design
Discover how Figma is transforming digital design with AI. David Kossnick sha...
Customizable, no-code voice agent automation with GP...
Retell AI is transforming the call center with AI voice automation powered by...
Toward understanding and preventing misalignment gen...
We study how training on incorrect responses can cause broader misalignment i...
New in ChatGPT for Business: April 2025
Watch hands-on demos of the lastest in ChatGPT for Business: o3, image genera...
Introducing OpenAI o3 and o4-mini
Our smartest and most capable models to date with full tool access
OpenAI o3 and o4-mini System Card
OpenAI o3 and OpenAI o4-mini combine state-of-the-art reasoning with full too...
Addendum to OpenAI o3 and o4-mini system card: OpenA...
We are replacing the existing GPT-4o-based model for Operator with a version ...
Addendum to o3 and o4-mini system card: Codex
Codex is a cloud-based coding agent. Codex is powered by codex-1, a version o...
The San Antonio Spurs use ChatGPT to scale impact on...
Discover how the San Antonio Spurs are using custom GPTs to enhance fan engag...
Sycophancy in GPT-4o: what happened and what we’re d...
We have rolled back last week’s GPT‑4o update in ChatGPT so people are now us...
