OpenAI has rolled out an update to its latest AI model, GPT-5, making it “warmer and friendlier” after some users complained it felt too blunt compared to earlier versions.
The highly anticipated GPT-5 launched earlier this year, but CEO Sam Altman admitted the rollout was “a little more bumpy than we’d hoped for.” Many users voiced a preference for GPT-4o, calling its tone more approachable.
In response, OpenAI says it has made subtle adjustments to GPT-5’s personality, aiming to make interactions feel smoother and more natural.
What’s Changing in GPT-5
According to the company, the update adds small conversational touches like “Good question” or “Great start” without resorting to excessive flattery.
OpenAI emphasized that internal testing shows no increase in sycophancy compared to the previous GPT-5 version, suggesting the model will remain concise while also being more human-like in its responses.
Executives Hint at Broader Plans
At a recent dinner with journalists, OpenAI executives tried to shift the conversation toward future projects beyond GPT-5. However, the rocky launch of the new model dominated the discussion.
VP Nick Turley acknowledged that GPT-5 initially felt “just very to the point,” but stated that the latest update aims to make it “more approachable now.”
OpenAI hopes these personality tweaks will reassure users who found GPT-5 less engaging than earlier models. This update marks a step toward balancing precision with approachability as the company looks ahead to its next AI innovations.
