GPT-5.1 Arrives: OpenAI Unveils ‘Adaptive Reasoning’ & New Human-Like Chat Styles

OpenAI has announced GPT-5.1, a major revision within the GPT-5 generation that aims to make ChatGPT more intelligent and more human in conversation, with two new models—GPT-5.1 Instant and GPT-5.1 Thinking.

According to OpenAI, GPT-5.1 Instant adopts a warmer, more conversational tone and introduces adaptive reasoning to decide when deeper thinking is needed, with reported gains in the AIME 2025 mathematics competition and Codeforces programming evaluations; GPT-5.1 Thinking is a more precise reasoning model that allocates thinking time differently for simple versus complex tasks.

The update also expands ChatGPT’s personalisation, adding three new preset tones—Professional, Candid, and Quirky—and experimental fine-tuning to adjust brevity, enthusiasm, scannability, and emoji frequency across all conversations. OpenAI says GPT-5.1 Instant and Thinking will begin rolling out to all users starting 12 November 2025, with paid subscribers receiving priority access and Enterprise and Education customers granted a seven-day early-access period.

After deployment, GPT-5.1 will become ChatGPT’s new default model, while the original GPT-5 models will remain available to paying subscribers for three months, and API support is expected later this week.