OpenAI users have reported significant performance issues with the GPT-5.5 model, alleging that the model has been silently downgraded to a less capable version. Users paying $200 monthly for the service have expressed frustration over what they describe as "fake thinking," where the model's interface still displays "GPT-5.5 Extended Thinking," but the quality of responses has deteriorated. Complaints have surfaced on platforms like the OpenAI developer forum and social media, with users noting that the model's performance now resembles that of earlier versions. The controversy intensified after users discovered that once their message quota is exhausted, the system automatically switches to a lower-tier model without notification. This "silent" downgrade has led to widespread dissatisfaction, with users feeling misled about the service they are receiving. OpenAI's official documentation confirms this practice, describing it as a feature, which has only fueled further criticism. Despite these issues, reports indicate that OpenAI is already testing a new version, GPT-5.6, in the background, even as users struggle with the current model's performance.