GPT-5.3 Instant: Smoother, more useful everyday conversations

OpenAI
GPT-5.3 Instant improves ChatGPT's conversational flow, reduces unnecessary refusals, enhances web search synthesis, and delivers more accurate responses.

Summary

OpenAI has released GPT-5.3 Instant, an update to ChatGPT's most-used model, focusing on making everyday conversations more fluid and helpful by addressing tone, relevance, and flow based on user feedback. Key improvements include significantly reducing unnecessary refusals and overly cautious disclaimers, leading to more direct answers. The model now synthesizes web search results more effectively, balancing external information with internal knowledge to provide relevant, upfront answers without over-indexing on links. Conversationally, it adopts a smoother, more to-the-point style, cutting out phrases that felt 'cringe' or overbearing. Furthermore, GPT-5.3 Instant shows improved factual accuracy, reducing hallucinations by up to 26.8% in high-stakes domains when using the web compared to prior models. It is also a stronger writing partner, capable of more resonant prose. Limitations remain in non-English language naturalness, but the model is available immediately to all ChatGPT users and developers via the API as 'gpt-5.3-chat-latest,' with GPT-5.2 Instant remaining available for three months.

(Source:OpenAI)