GPT-5 gets a little update

Though we may not have reached the heights we anticipated yesterday, today is a brand new day to begin a new climb.

GPT-5 gets a little update

OpenAI just made a small but meaningful update to GPT-5, designed to make it feel friendlier when you chat with it. After the model first launched, many users said it came across as too blunt and a little cold compared to earlier versions. Even CEO Sam Altman admitted the rollout was “a little more bumpy than we’d hoped for.”

The new changes add simple touches like “Good question” or “Great start” at the right moments. The idea isn’t to flatter or overdo it, but to make conversations feel more natural and human. OpenAI says the update doesn’t make the model more sycophantic, just a bit warmer and easier to talk to. This shows that the company is paying attention to feedback and fine-tuning how the AI interacts, not just how it performs. It may sound like a small detail, but for people who use GPT-5 for teaching, support, or daily work, even a shift in tone can make the whole experience feel more engaging and approachable.