OpenAI releases an upgraded version of GPT-5.1, which enhances and strengthens "adaptive reasoning" and dialogue style through the Instant and Thinking dual models.

OpenAI announces the launch ofGPT-5.1 version updateThis upgrade represents a major overhaul of the GPT-5 generation, with the core goal of making ChatGPT not only smarter but also more human-like in its conversational interactions. OpenAI emphasizes that GPT-5.1 will achieve meaningful improvements in both intelligent performance and communication style through two new models—GPT-5.1 Instant and GPT-5.1 Thinking.

Instant introduces adaptive inference, making Thinking more responsive.

The two core models updated this time each have their own positioning:

• GPT-5.1 Instant:As the most widely used model in ChatGPT, the new version of Instant's preset tone has become warmer and more talkative, adding a touch of playfulness while maintaining its practicality.

More importantly, it introduces an "adaptive reasoning" mechanism for the first time. The model can automatically determine when more time needs to be spent thinking about a difficult problem, which the official team claims has led to significant improvements in the AIME 2025 math competition and Codeforces programming evaluation.

• GPT-5.1 Thinking:This is OpenAI's advanced reasoning model. The new version is more precise in allocating thinking time, reacts more quickly to simple tasks, and is more persistent in handling complex problems. In addition, its responses have reduced the use of obscure technical jargon, making it easier for users to understand complex technical concepts.

This is an advertisement.

Enhanced style customization: New professional, frank, and quirky modes added.

In addition to improvements in the model's intelligence, OpenAI has also significantly updated ChatGPT's personalization settings. The official documentation introduces more intuitive control options, allowing users to more easily shape the AI's tone of voice.

• New preset styles:In addition to the existing "Default", "Friendly", and "Efficient" styles, three new styles have been added: "Professional", "Candid", and "Quirky".

• Experimental fine-tuning:Some users will be able to make more subtle adjustments to the conciseness, enthusiasm, scannability, and frequency of emoji usage in their replies.

• Effective immediately:All settings changes will now apply immediately to all conversations (including ongoing ones) without needing to start a new chat.

Launch timeline: Paid users will get early access, with the API launching later this week.

OpenAI announced that GPT-5.1 Instant and Thinking will begin rolling out to all users starting November 12th (US time). Paid users (including Pro, Plus, Go, and Business plans) will be the first to experience it, while Enterprise and Edu users will have a 7-day early bird access period.

Following this, GPT-5.1 will officially become the default model for ChatGPT. To ensure a smooth transition, existing GPT-5 models (including Instant and Thinking models) will be retained in the old model menu for paid subscribers for three months. As for API support, which developers are concerned about, it is expected to be released later this week.

This is an advertisement.

This is an advertisement.

Tags: AIChatGPTGPT-5GPT-5.1LLMOpenAIArtificial wisdom