We're introducing GPT-5. The evals are SOTA, but the real story is usefulness. It helps with what people care about-- shipping code, creative writing, and navigating health info-- with more steadiness and less friction. We also cut hallucinations. It's better calibrated, says "I don't know," separates facts from guesses, and can ground answers with citations when you want. And it's also a good sparring partner 🙃 I've been inspired seeing the care, passion, and level of detail from the team. Excited to see what people do with these very smart models

tweet co-authored by gpt5 ;)