From hard refusals to safe-completions: toward output-centric safety training
August 7, 2025 SafetyRelease Introduced in GPT‑5, safe-completion is a new safety-training approach to maximize model helpfulness within safety…
Kategória: OpenAI
Introducing GPT-5
August 7, 2025 ReleaseProduct Our smartest, fastest, most useful model yet, with built-in thinking that puts expert-level intelligence in…
Introducing GPT-5 for developers
August 7, 2025 Product The best model for coding and agentic tasks. Loading… Introduction Today, we’re releasing GPT‑5 in…
GPT-5 and the new era of work
August 7, 2025 Product Loading… Today we’re introducing GPT‑5, OpenAI’s smartest, fastest, most useful model yet, and a major…
Providing ChatGPT to the Entire U.S. Federal Workforce
August 6, 2025 CompanyGlobal Affairs First-of-its-kind partnership with Government Services Administration will give federal agencies access to ChatGPT Enterprise…
Estimating worst case frontier risks of open weight LLMs
August 5, 2025 SafetyPublication Read the paper(opens in a new window) Abstract In this paper, we study the worst-case…
gpt-oss-120b & gpt-oss-20b Model Card
August 5, 2025 PublicationSafety Read the model card(opens in a new window) Introduction We introduce gpt-oss-120b and gpt-oss-20b, two…
Open Weights and AI for All
August 5, 2025 Global Affairs Loading… AI’s next frontier isn’t just about capability—it’s about who gets to use it.…
Introducing gpt-oss
August 5, 2025 ReleaseProduct gpt-oss-120b and gpt-oss-20b push the frontier of open-weight reasoning models Explore on Hugging Face(opens in…
What we’re optimizing ChatGPT for
We build ChatGPT to help you thrive in all the ways you want. Learn how we’re improving support for…