OpenAI Launches GPT-5.5 Instant as ChatGPT Default

OpenAI has replaced ChatGPT's default model with GPT-5.5 Instant, promising fewer hallucinations, greater accuracy, and more concise responses. The update targets sensitive topics like finance, law, and healthcare, with inaccuracies on flagged conversations reduced by 37.3%. The model also cuts back on excessive emojis and unnecessary follow-up questions. The new model improves on benchmark scores from its predecessor, GPT-5.3 Instant, including a jump from 32.9 to 38.4 on HealthBench Professional. It also draws more actively from users' previous chats and connected accounts for personalized responses, while giving users full control over their stored context. The rollout is global starting today.