OpenAI has announced an iterative update to GPT-5, dubbed GPT-5.1, which includes two models — GPT-5.1 Instant and GPT-5.1 Thinking — both designed to make ChatGPT more adaptable, easier to use, and noticeably better at understanding and responding to different kinds of questions.
GPT-5.1 Instant is the model most people will interact with, and it now sounds warmer, smarter, and more natural and conversational. It follows instructions more accurately and uses adaptive reasoning to spend extra time on tougher questions without slowing down simple chats. Early testing by OpenAI shows clear improvements in math and coding tasks, including exams like AIME 2025 and platforms such as Codeforces.
GPT-5.1 Thinking focuses on complex reasoning. It adjusts its response time depending on how difficult the query is — offering quicker answers for straightforward questions and more detailed explanations for challenging ones. Compared to the previous GPT-5 Thinking model, it’s about twice as fast with easy tasks and up to twice as slow on the hardest ones, ensuring deeper, more thoughtful responses. The tone has also been refined to sound clearer, less technical, and more empathetic.
GPT-5.1 introduces several useful new features as well. Adaptive Model Routing automatically decides which model is better suited for each query. Users can pick from six chat styles — Default, Friendly, Efficient, Professional, Candid, and Quirky — while the older Cynical and Nerdy options still remain. There are new fine-tuning sliders to adjust warmth, conciseness, scannability, and Emoji use.
Also Read: OpenAI Debuts ChatGPT Atlas Browser with ChatGPT Built-In
ChatGPT can now proactively suggest tone or style changes during conversations, and any personalization you set takes effect instantly across all chats. The models also follow custom instructions more accurately, whether you’re using Instant or Thinking.
OpenAI says GPT-5.1 will continue to evolve, with upcoming updates focused on better reasoning, smoother conversations, and more powerful personalization tools. The goal is to make ChatGPT feel increasingly tailored to each individual user.
GPT-5.1 Instant and GPT-5.1 Thinking are currently rolling out to paid users — including Pro, Plus, Go, and Business plans — with free and logged-out users getting access next. Enterprise and Education plans have a seven-day early-access option. Over time, GPT-5.1 will become the default model for everyone, though the older GPT-5 models will remain available to paid users for the next three months.