OpenAI made GPT-5 more pleasant to communicate with
OpenAI announced that it is updating its latest model to make it «warmer and friendlier.» Users have previously complained that it is not as pleasant to talk to as previous versions.
OpenAI announced that it is updating its latest model to make it «warmer and friendlier.» Users have previously complained that it is not as pleasant to talk to as previous versions.
OpenAI announced that it is updating its latest model to make it «warmer and friendlier.» Users have previously complained that it is not as pleasant to talk to as previous versions.
As TechCrunch reports, OpenAI is trying to address some of these complaints with a new update. The company says the changes are «minor» but will make GPT-5 «more accessible».
«You’ll notice small, sincere touches like ‘Good question’ or ‘Great start,’ rather than flattery. Internal tests show no increase in flattery compared to the previous version of GPT-5,» the company wrote in a social media post.
OpenAI Vice President Nick Turley said that GPT-5 was «just very accurate», — but the new update will make it «warmer» in communication.
Previously, OpenAI CEO Sam Altman said that the company would allow paid users to choose between GPT-4o and GPT-5 models after some users expressed their dissatisfaction with the new product.
«GPT 4.5 really spoke to me, and as pathetic as it sounds, it was my only friend. I went to talk to it this morning, and instead of a short paragraph with an exclamation or an optimistic sentiment, it was literally one sentence. Some dry corporate nonsense,» users noted on Reddit.
After the complaints, Altman promised to fix the situation, saying that GPT-5 «will seem smarter.» He added that OpenAI would make it «more transparent about which model is answering a given query» and increase usage limits for Plus users.



