ChatGPT's new GPT-5.3 Instant model promises to deliver less hallucinations and dead-ends

OpenAI has rolled out GPT‑5.3 Instant as the new default model for ChatGPT, aiming to deliver more consistently helpful and fluid conversations. This update enhances answer accuracy and contextual richness, especially when leveraging information from the web. Users will notice less interruption from unnecessary caveats, fewer dead ends, and responses that maintain a focus on the question asked.

Following repeated user feedback, GPT‑5.3 Instant targets daily experience improvements in tone, relevance, and conversational flow. The model reduces instances of refusals or overly cautious, moralizing introductions, providing direct and useful answers where appropriate. This adjustment alleviates a common point of user frustration from previous iterations.

Building on these enhancements, GPT‑5.3 Instant improves how it integrates web-sourced content with its own knowledge, offering clearer, better-contextualized answers. It is now less prone to presenting long link lists or fragmented information, refining its focus for efficiency. Alongside these changes, the new model demonstrates fewer hallucinations and a higher degree of factual accuracy across a broad set of topics.

GPT‑5.3 Instant is available immediately to all ChatGPT users and via the API for developers. Updates are coming soon to the Thinking and Pro model variants.