ChatGPT can now proactively suggest updating settings during a conversation when the user inquires about the style or tone of communication. There is also the option for fine-tuning parameters: adjusting the conciseness, warmth, and comfort of responses, as well as the frequency of emoji usage. By default, GPT-5.1 Instant is warmer, more playful, and adheres to the specified parameters more frequently, while GPT-5.1 Thinking can more accurately adapt its thinking time to the question, making responses clearer and less ambiguous. At the same time, it remains warmer and more empathetic than before.

User questions will be automatically directed to the most appropriate model — GPT-5.1 Auto. The rollout of the Instant and Thinking models will begin gradually. Initially, access will be granted to paid users of Pro, Plus, Go, and Business, followed by free and unregistered users. Enterprise and Edu users will be able to test the new models for seven days, after which GPT-5.1 will become the default standard model.

OpenAI plans a gradual rollout of GPT-5.1 to ensure stable system performance. As a result, for the next three months, GPT-5 will remain available to paid subscribers in legacy models.