OpenAI has released GPT-5.3 Instant, a major update to ChatGPT's most-used model that delivers smoother, more natural everyday conversations. The update focuses on three core improvements: more accurate answers, better web search integration, and a conversational style that cuts down on unnecessary disclaimers and overly cautious responses. GPT-5.3 Instant directly addresses user feedback about frustrating dead ends and stilted interactions.

This release represents OpenAI's most comprehensive effort yet to refine the practical ChatGPT experience. The model now provides more factual responses with hallucination rates dropping by 22.5% on user-flagged errors and 26.8% in higher-stakes domains when using web search. It also better balances online information with its own knowledge, surfacing the most relevant details upfront rather than overwhelming users with link lists or loosely connected data.

GPT-5.3 Instant also improves writing capabilities, helping users craft more resonant prose for creative projects. The update is available immediately to all ChatGPT users and developers via the API as 'gpt-5.3-chat-latest.' OpenAI plans to retire GPT-5.2 Instant for paid users on June 3, 2026, after a three-month transition period. While the model shows significant progress in English conversations, OpenAI acknowledges ongoing work needed for non-English languages and continued tone refinement.