With OpenAI’s latest model GPT-5.5 now released, users who missed GPT-4o, which ended in February, are voicing expectations and assessments on both performance and emotional appeal.

According to IT outlet Business Insider on May 1 local time, GPT-4o was widely favored for lively, distinctive responses, and some users viewed it like a friend or adviser. But OpenAI judged the model to have an overly flattering tendency and adjusted its personality. The GPT-5.0 and GPT-5.2 models that followed were designed to deliver more restrained responses.

In the process, the latest models reduced flattering behavior but were also criticized for weaker flexibility and expressiveness. Some users said interactions felt more rigid than before and that the models took a more cautious and limited stance even in providing information. Some also said questions involving medical issues or advice drew less proactive interpretation or fewer suggestions than in the past.

By contrast, assessments of GPT-5.5 have been relatively positive. Users say the new model carries conversational context more naturally and presents views steadily across a range of topics. Online communities that had called for GPT-4o’s return also show reactions that 5.5 has improved compared with earlier models.

Still, some users say they remain disappointed on emotional elements. Some users said GPT-5.5 has advanced functionally but has not fully recreated the intuitive, familiar interaction experience that GPT-4o delivered. Separately, OpenAI said it acknowledged and fixed an issue in the 5.1 model in which certain personality settings triggered excessive expression.