OpenAI promises more warmth and personality with its new GPT-5.1

Just a heads up, if you buy something through our links, we may get a small share of the sale. It’s one of the ways we keep the lights on here. Click here for more.

OpenAI is back with another sequel, GPT-5.1, because in the year 2025, even AI models apparently need mid-season refreshes. 

Dropping today, the update aims to make ChatGPT “smarter and more enjoyable to talk to,” which is Silicon Valley–speak for “please stop tweeting that our last release was mid.”

GPT-5.1 comes in two flavors: Instant and Thinking. Instant is the breezier sibling, “warmer, more intelligent, and better at following your instructions,” according to OpenAI.

Think of it as the model that actually listens when you say, “give me bullet points.” Then there’s Thinking, which has been tuned to be “easier to understand and faster on simple tasks, and more persistent on complex ones.” 

So it’s basically the studious one who stays late after class. ChatGPT will auto-route your questions to whichever version makes the most sense, like a digital maître d’ deciding which waiter can handle your order of emotional support paragraphs.

The rollout begins this week, and the older GPT-5 models will live in the legacy menu for about three months before being escorted off the premises.

OpenAI is also expanding its growing wardrobe of personality presets, because apparently, 800 million ChatGPT users all want their AI to sound slightly different. 

The lineup now includes Default, Professional, Friendly, Candid, Quirky, Efficient, Nerdy, and Cynical. (Good luck resisting the urge to click Cynical.) 

The company is also testing an experimental setting that lets users fine-tune ChatGPT’s style directly, basically a mood slider for your chatbot.

Fidji Simo, OpenAI’s CEO of Applications, summed it up in a Substack post: with hundreds of millions of users, “we’re well past the point of one-size-fits-all.” 

In other words, ChatGPT is now a customizable digital roommate.

All this comes after GPT-5’s underwhelming debut in August, which left users unimpressed enough that OpenAI had to resurrect GPT-4o on the very next day. 

Even Microsoft has been quietly shopping around, turning to Anthropic’s models for Copilot Researcher, GitHub Copilot, Copilot Studio, and a new Office Agent.

GPT-5.1 lands just weeks after OpenAI launched its AI-powered browser, ChatGPT Atlas, complete with an “agent mode” that can click around the web for you like a particularly obedient intern.

Whether GPT-5.1 is the redemption arc fans wanted or just another patch, rollout begins now.