ChatGPT New GPT-5.3 Instant Model Aims to Improve “Everyday Usability”

OpenAI released today a new GPT-5.3 Instant model to address the main shortcomings of its existing GPT-5.2 Instant model. The company acknowledged that GPT‑5.2 Instant could sometimes refuse to answer “safe” questions and respond in an overly defensive or “cringe” manner, and this new model should help to make everyday conversations more seamless.

“This update focuses on the parts of the ChatGPT experience people feel every day: tone, relevance, and conversational flow. These are nuanced problems that don’t always show up in benchmarks, but shape whether ChatGPT feels helpful or frustrating. GPT‑5.3 Instant directly reflects user feedback in these areas,” the company said today.

In practice, ChatGPT users should notice that the new model “significantly reduces” unnecessary refusals and responds with a more to-the-point style. The quality of responses grounded in web information should also be improved, with GPT-5.3 doing a better job at synthesizing information. Users should also experience fewer hallucinations compared to previous models.

Lastly, GPT-5.3 should improve ChatGPT’s writing skills. “It’s better at helping you write resonant, imaginative, and immersive prose, whether you’re drafting fiction, refining a passage, or exploring new ideas,” the company said today.

GPT‑5.3 Instant is rolling out today to all ChatGPT users, and OpenAI plans to also update its GPT-5.2 Thinking and Pro variants soon. GPT‑5.2 Instant will remain available for ChatGPT paid users until June 3, 2026.