OpenAI launches GPT-5.1 with more natural responses

OpenAI is beginning to roll out GPT-5.1, an update to the GPT-5 family designed to improve two things at once: its ability to solve tasks and the way it holds a conversation. The rollout starts with paid users and will gradually reach everyone else.

The update includes two models:

• GPT-5.1 Instant, designed to respond quickly in everyday use.
• GPT-5.1 Thinking, built for problems that require more reasoning.

The company is also adding new options to adjust ChatGPT’s tone and make its responses better suited to each person.

A clearer, less rigid ChatGPT

GPT-5.1 Instant will be the most widely used model in ChatGPT, and it now has a warmer, more conversational tone by default. OpenAI says it also follows instructions more closely, so it should respond more precisely to what you are actually asking instead of drifting into unnecessary explanations.

The goal is not simply to make ChatGPT friendlier. The idea is to help it communicate more naturally without losing clarity or usefulness. In practice, it should work just as well for drafting an email in a specific tone as for explaining a concept without sounding so mechanical.

For the first time, GPT-5.1 Instant includes adaptive reasoning. This means it decides when it needs to spend more time thinking before answering. It responds quickly to a simple question, but it can pause longer and produce a more complete answer for a math or programming problem.

OpenAI points to improvements in evaluations such as AIME 2025, which focuses on mathematics, and Codeforces, which is related to programming. The company does not provide a single figure in this announcement summarizing how much it improves on each test.

Thinking spends time based on the task

GPT-5.1 Thinking also adjusts its reasoning time based on difficulty. Simple questions should take less time, while complex tasks receive a more persistent analysis.

Its responses are also designed to be easier to understand. The model reduces jargon and unexplained terms, which is especially useful when you use it to analyze work documents, solve technical problems, or learn a new subject.

Its default tone will also be warmer and more empathetic. ChatGPT Auto will continue choosing the most suitable model for each query automatically, so in most cases you will not have to decide which one to use.

More control over personality

GPT-5.1 comes with new style settings. In addition to the existing options, ChatGPT now includes these profiles:

• Default, the general style.
• Friendly, previously called Listener.
• Efficient, previously called Robot.
• Professional, for more formal responses.
• Candid, more direct.
• Quirky, with a more distinctive tone.

Cynical and Nerdy will remain available. These settings work with all models, not just GPT-5.1.

OpenAI is also testing more precise controls for choosing how much ChatGPT should summarize, how warm or easy to scan it should be, and how often it should use emojis. In some cases, ChatGPT may suggest a style change if it detects that you are repeatedly asking for a particular type of response.

One important change is that personalization updates apply immediately to all chats, including conversations that were already open. Previously, some settings only affected new conversations.

When it will be available

The rollout starts with the Pro, Plus, Go, and Business plans, then will reach free users and people who are not signed in. It will happen gradually over the next few days, so it will not necessarily appear in your account right away.

Enterprise and Edu users will get seven days of early access through an option that will be disabled by default. After that, GPT-5.1 will become the default model.

Paid subscribers can continue using the current GPT-5 models for three months from the legacy models menu. This will make it possible to compare results before adapting to the change. GPT-5 Pro will also be updated soon to GPT-5.1 Pro.

The update will reach the API later during the same week. GPT-5.1 Instant will be available as gpt-5.1-chat-latest, while GPT-5.1 Thinking will be offered as GPT-5, with both using adaptive reasoning.

For you, the most visible change will not necessarily be a much longer response or a new standalone feature. It will be that ChatGPT understands what you are asking more accurately, takes less time on simple tasks, and maintains a more natural, configurable tone. What matters now is whether those improvements hold up in everyday use, especially when questions are ambiguous or require real work.