• Source:JND
Add as a preferred

Weeks after the rocky debut of its latest flagship AI model, OpenAI says it is adjusting GPT-5 to make its answers sound friendlier and less stiff.

The change follows user complaints that GPT-5’s responses “felt too formal.” In a post on X, the company explained, “Changes are subtle, but ChatGPT should feel more approachable now. You’ll notice small, genuine touches like ‘Good question’ or ‘Great start’, not flattery. Internal tests show no rise in sycophancy compared to the previous GPT-5 personality.”

The update will begin rolling out this week. It’s part of a series of changes OpenAI has announced since GPT-5 went live on August 7. The release, hyped since GPT-4 in 2023, left many underwhelmed when it finally arrived after delays caused by safety reviews and compute limits.

ALSO READ: Google Pixel 10 Pro Fold Design And Key Features Revealed In Leaks

When GPT-5 became available to all ChatGPT users this month, many noted that the improvements felt modest, with the biggest upgrades tied to cost and speed. At a dinner with journalists, OpenAI CEO Sam Altman reportedly admitted, “I think we totally screwed up some things on the rollout.”

Still, Altman pointed to a surge in adoption: “On the other hand, our API traffic doubled in 48 hours and is growing. We’re out of GPUs. ChatGPT has been hitting a new high of users every day. A lot of users really do love the model switcher. I think we’ve learnt a lesson about what it means to upgrade a product for hundreds of millions of people in one day.”

ChatGPT’s weekly user base has now climbed to nearly 700 million, quadrupling in a year.

Nick Turley, ChatGPT’s product head, also acknowledged the feedback on GPT-5’s tone. “GPT-5 was just very to the point. I like that. I use the robot personality — I’m German, you know, whatever. But many people do not, and they really like the fact that ChatGPT would actually check in with you.”

New GPT-5 modes arrive

Among the biggest changes are new operating modes: Auto, Fast, and Thinking. Fast mode delivers quicker responses, while thinking mode takes longer to generate more in-depth answers. Auto switches between the two. Users can pick modes directly within ChatGPT.

Altman highlighted the importance of this feature: “GPT-5 will seem smarter starting today. Yesterday, the autoswitcher broke and was out of commission for a chunk of the day, and the result was GPT-5 seemed way dumber.”

GPT-4o returns after user push

Another key update comes from user feedback. During a Reddit AMA, many requested the return of GPT-4o. Altman confirmed that the model is now back under “Legacy models” for Plus subscribers, along with O3, GPT-4.1, and GPT-5 Thinking Mini. To access them, paid users need to enable “Show additional models” in settings.

Altman also promised more transparent “transition periods” when retiring models in the future.

ALSO READ: Samsung Galaxy S25 FE Renders Leak, Reveal Colours And Flat-Frame Design

Higher limits for subscribers

OpenAI has raised usage caps as well. Plus and Team subscribers can now send up to 3,000 weekly messages using GPT-5 in Thinking mode, with extra access on GPT-5 Thinking mini if that limit is exceeded. GPT-5 is also now available for Enterprise and Edu plans.

The company is further refining ChatGPT’s interface to make it easier to enable Thinking Mode and to clearly show which model is responding.