Just as the online rumors from a few days ago suggested,OpenAI Released the GPT-5.1 model, an upgraded version of the original GPT-5. In addition to being smarter and smoother in conversation, it also offers more customizable style and tone features, allowing you to set more suitable ones based on different chat content. AI 的回答風格。

OpenAI launches new GPT-5.1 model with customizable style and tone, smarter and more conversational.
Earlier, OpenAI announced on its official blog that the new GPT-5.1 model is about to launch, and starting today it will begin rolling out to everyone in phases—initially for paid users, while free users will need to wait a bit longer.
Like GPT-5, this time GPT-5.1 will also offer two models, “GPT-5.1 Instant” and “GPT 5.1 Thinking” — the former is suitable for most purposes, while the latter is suited for complex tasks:
- GPT-5.1 Instant: The most versatile model, now warmer, smarter, and better at following your instructions by default.
- GPT-5.1 Thinking: Advanced reasoning model, now easier to understand, faster at simple tasks, and more resilient on complex ones.

The highlight this time is “GPT-5.1 Instant.” OpenAI says it has added an “adaptive reasoning” capability for the first time: when faced with harder questions, the AI automatically decides whether to think deeply first, making responses more comprehensive and accurate while maintaining fast response times.
GPT-5 Thinking is more efficient in everyday use, adjusting thinking time more accurately based on the question: more complex questions take more time, while simpler ones get faster answers.
Compared to GPT-5, GPT-5.1 Thinking is about twice as fast on simple tasks, but about twice as slow on complex tasks (spending more time on deep thinking), making it more efficient overall.

In addition to the new model, OpenAI has also made it easier to customize ChatGPT’s tone and style, redesigning the tone options. The original Default, Friendly (formerly Listener), and Efficient (formerly Robot) are still available, with new additions Professional, Candid, and Quirky:

Also testing more granular customization options—in the future, you’ll be able to adjust ChatGPT’s tone and style in settings, including: length, tone temperature, how quickly it can be skimmed, and emoji usage frequency.
GPT-5.1 Instant and Thinking will begin rolling out to paid users (Pro, Plus, Go, Business) starting today, before being made available to free and logged-out users. Enterprise and Edu plans can switch to Early-Access within the next seven days (off by default).
And the older GPT-5 (Instant and Thinking) will remain in ChatGPT for three months,
Therefore, if you’re not used to it, you can temporarily continue using GPT-5, but note that it will still be discontinued after three months. The other old models are unaffected.
OpenAI also previewed that later this week, GPT-5.1 Instant will launch on the API as gpt-5.1-chat-latest, and GPT-5.1 Thinking will launch on the API as GPT-5.1, both with adaptive reasoning capabilities.
Source: KOCPC Chinese