OpenAI has announced an update to its latest model, GPT-5, aimed at making the system “warmer and friendlier.” The update was disclosed late Friday, as the company continues to refine the user experience following a somewhat rocky rollout.

The launch of GPT-5, which has been highly anticipated, was described by CEO Sam Altman as “a little more bumpy than we’d hoped for.” Some users had expressed a preference for the earlier model, GPT-4o, citing that GPT-5 felt less approachable.

In response, OpenAI has introduced subtle adjustments to the model’s tone. The company explained that users may now encounter phrases such as “Good question” or “Great start” to make interactions feel more natural. Importantly, OpenAI stressed that these changes are not designed to flatter, and internal testing revealed no increase in sycophancy compared to the earlier GPT-5 personality.

At a recent dinner with journalists, OpenAI executives sought to shift the discussion toward the company’s broader plans beyond GPT-5. However, the challenges surrounding the launch remained a central talking point.

Vice President Nick Turley acknowledged that the original version of GPT-5 was “just very to the point,” but emphasized that the latest update would make it “feel warmer,” aligning with the company’s ongoing effort to balance precision with approachability.

Source: Tech Crunch