OpenAI launches GPT-5.1, a smarter and more conversational AI model
What's the story
OpenAI has unveiled a major upgrade in its GPT-5 series with the launch of GPT-5.1. The company describes the new model as "smarter, more reliable, and a lot more conversational." OpenAI CEO Sam Altman took to X to share his excitement about the update, particularly highlighting improvements in instruction following and adaptive thinking capabilities.
Model variants
Two new models: GPT-5.1 Instant and GPT-5.1 Thinking
OpenAI is expanding the GPT-5 series with two new models: GPT-5.1 Instant and GPT-5.1 Thinking. The former is described as "warmer, more intelligent, and better at following your instructions," while the latter is an advanced reasoning model that excels at simple tasks and complex ones alike. The rollout will start with paid users on Pro, Plus, Go, and Business plans before reaching free users later on.
User control
Preset personality options get an upgrade
OpenAI has also enhanced ChatGPT's preset personality options, giving users more control over its responses. The existing modes, Default, Friendly (formerly Listener), and Efficient (formerly Robot), will be updated with three new tones: Professional, Candid, and Quirky. These settings will apply across all models. For finer control, users can directly adjust ChatGPT's behavior in the settings menu to suit their preferences for conciseness or warmth of replies.
Enhanced capabilities
Improvements in math and coding tasks
The GPT-5.1 model shows significant improvements in math and coding tasks, outperforming on AIME 2025 and Codeforces benchmarks. It also prioritizes clarity by avoiding unnecessary technical jargon and explaining concepts in simple terms. The rollout will start with paid users but enterprise and Edu plan holders will get a seven-day early access option (off by default). After that, GPT-5.1 will become the default model for all users.