OpenAI Wednesday released GPT-5.1 Instant and GPT-5.1 Thinking are two updated versions of flagship AI models that are now available on ChatGPT. The company wraps the models in the language of anthropomorphism, claiming they are friendlier, more talkative and better at following instructions.
The release follows complaints earlier this year that previous models were too cheerful and sycophanticas well as controversial differences among users over how OpenAI modified default GPT-5 output style after several suicide lawsuits.
Now the company is facing close attention from lawyers and regulators that could jeopardize its future operations. In this environment, it's difficult to simply release a new AI model, add some statistics and move on, as a company could even a year ago. But here are the basics: the new GPT-5.1 Instant model will serve as the default ChatGPT option for most tasks, while the GPT-5.1 Thinking model will simulated reasoning model which attempts to cope with more complex problem solving tasks.
OpenAI claims that both models perform better on technical tests such as calculus and coding assessments (including AIME 2025 and Codeforces) than GPT-5which was released in August.
Improved benchmarks may appeal to some users, but the biggest change in GPT-5.1 is its presentation. OpenAI says it has heard from users that they want AI models to simulate different communication styles depending on the task, so the company offers eight preset options, including Professional, Friendly, Frank, Quirky, Efficient, Cynical and Crazy, as well as a default setting.
These presets change the instructions given at each prompt to model different personality stylesbut the base model's capabilities remain the same across all settings.
The company also trained GPT-5.1 Instant to use “adaptive reasoning,” which means the model decides when to spend more computational time processing a hint before generating output.
The company plans to gradually roll out these models over the next few days, starting with paid subscribers and then expanding to free users. Later this week, OpenAI plans to add GPT-5.1 Instant and GPT-5.1 Thinking to its API. GPT-5.1 Instant will be exposed in the API as gpt-5.1-chat-latest, and GPT-5.1 Thinking will be released as GPT-5.1 in both cases with adaptive reasoning enabled. Legacy GPT-5 models will remain available on ChatGPT in the legacy models dropdown for paying subscribers for three months.
