OpenAI says ChatGPT is now smarter, warmer
Add Axios as your preferred source to
see more of our stories on Google.
OpenAI
OpenAI is updating ChatGPT to have a more conversational tone, while also making it work faster for simple tasks and harder and longer for complex ones.
Why it matters: Users have formed strong attachments to ChatGPT's personality and have openly rebelled when it changes.
- OpenAI says the updates are designed to make ChatGPT "more enjoyable to talk to."
Driving the news: The updates to GPT-5 make ChatGPT communicate in a more natural way.
- The company previously allowed users to customize bots' responses with presets including "Cynic," "Robot," "Listener" and "Nerd."
- The GPT-5.1 update adds "Friendly," "Efficient," "Professional," "Candid," and "Quirky" and keeps the old tones.
Between the lines: Both versions of the model — 'Instant' for quick replies and 'Thinking' for complex reasoning — get an upgrade.
- GPT-5.1 Auto now chooses the right model automatically, so users usually don't have to.
What they're saying: OpenAI's CEO of applications, Fidji Simo, wrote in a blog post that while the company's mission is to build AI that benefits humanity, it also wants chatbots that are useful in daily life.
- "With more than 800 million people using ChatGPT, we're well past the point of one-size-fits-all," Simo writes.
- "We want ChatGPT to feel like yours and work with you in the way that suits you best," she adds.
Yes, but: Users and their loved ones already report forming unhealthy relationships with chatbots that can replace their real-world interactions, causing real psychological harm in some cases.
- "We have to be vigilant about the potential for some people to develop attachment to our models," Simo says, adding that the company is working closely with experts "to better understand what healthy, supportive interactions with AI should look like."
The fine print: GPT-5.1 starts with paid users and then will expand gradually to all.
- OpenAI says that the previous version will remain available in ChatGPT under the legacy models dropdown menu for paid subscribers for three months, "so people have time to compare and adapt at their own pace."
