GPT‑5.3 Instant Makes ChatGPT More Intelligent and Accurate

OpenAI has launched GPT‑5.3 Instant, an update designed to make everyday interactions more helpful, relevant, and natural. This version focuses on the nuances of conversational flow, tone, and contextual understanding — areas that strongly influence user experience but are not always captured by standard benchmarks.

The update reduces unnecessary refusals and overly cautious or moralizing responses, allowing the model to provide direct, focused answers. GPT‑5.3 Instant also improves web-based responses by synthesizing online information with its own knowledge, highlighting the most important points without relying on long lists of loosely connected links. Users can now expect answers that are more immediately useful, accurate, and relevant. Conversational style has been refined to feel smoother and more natural, avoiding overbearing or awkward phrasing. GPT‑5.3 Instant also strengthens its role as a writing assistant, capable of generating clear, imaginative, and immersive prose for both practical and creative tasks. 

Accuracy has improved significantly: in critical domains such as medicine, law, and finance, hallucination rates decreased by 26.8% when using the web and by 19.7% when relying solely on internal knowledge. In evaluations based on user feedback identifying factual errors, hallucinations dropped by 22.5% with web access and by 9.6% without it.

GPT‑5.3 Instant is available to all ChatGPT users, as well as to developers via the API as gpt-5.3-chat-latest. Updates for the Thinking and Pro versions will be released later. GPT‑5.2 Instant will remain temporarily available only to paid users in the Legacy Models section for three months before being retired on June 3, 2026.