Late Friday, OpenAI shared that it’s giving its latest model a small makeover—making it feel “warmer and friendlier.”
The rollout of GPT-5, which many had been eagerly waiting for, wasn’t exactly smooth sailing. CEO Sam Altman even admitted it was “a little more bumpy than we’d hoped,” and some users said they actually liked the older GPT-4o better.
To address that feedback, OpenAI is tweaking GPT-5 with subtle changes meant to make it more approachable. Think gentle touches like “Good question” or “Great start”—not empty compliments, but natural responses. The company also reassured people that internal tests show GPT-5 isn’t becoming overly flattering compared to earlier versions.
At a dinner with reporters this week, OpenAI leaders tried steering the conversation toward what’s coming after GPT-5. Still, as journalist Max Zeff put it, the awkward launch was hard to ignore. VP Nick Turley summed it up best: GPT-5 was “very to the point,” but with this update, it should finally feel a bit warmer.
