OpenAI has begun rolling out an update to its newly launched GPT-5 model aimed at making the system “warmer and friendlier,” according to a TechCrunch report published 17 Aug. 2025. The change follows a bumpier-than-expected debut that left some users preferring the earlier GPT-4o release. OpenAI says the revised version adds brief acknowledgements such as “Good question” without increasing flattery or sycophancy, based on internal testing. Chief Executive Officer Sam Altman acknowledged the model’s rocky start during a recent dinner with journalists, while Vice President Nick Turley said GPT-5 had been “very to the point” and would now feel more approachable. Early developer feedback highlighted improved efficiency, stronger willingness to disagree and better code-base integration compared with competing systems such as Anthropic’s Claude 3.5 Sonnet and Google’s Gemini 2.5. OpenAI continues to refine GPT-5 as the company shifts focus toward future releases.