Cursor 在未披露的情况下使用 Kimi K2.5（一个中国 AI 模型）——每位开发者都应该关注的原因

On March 19, Cursor announced the launch of Composer 2, boasting an impressive benchmark score of 61.7% on Terminal-Bench 2.0, outperforming Claude Opus 4.6 at a fraction of the cost. The announcement referred to it as their 'first continued pretraining run' and 'frontier-level coding intelligence.' The author, who had been using Cursor for months, shared the news with their team. However, less than 24 hours later, a developer named Fynn inspected Cursor's API traffic and discovered that the model ID in the API response was Kimi K2.5, a detail that was not mentioned in Cursor's announcement. This revelation led the author to question their trust in AI tool announcements, highlighting the need for developers to be more cautious and aware of transparency issues when using AI tools.