On April 22, @Moonshot Kimi released a notice regarding functional abnormalities that occurred during the launch of the new model Kimi K2.6, stating that all users would be compensated by resetting their quotas. Kimi K2.6 was officially launched and open-sourced on April 20, featuring comprehensive upgrades that offer stronger code writing capabilities, long-term task execution capabilities, and support for Agent clusters.
In the notice, Moonshot mentioned that the launch of Kimi K2.6 caused a surge in user traffic, leading to some users experiencing queueing and temporary unavailability of functions. Additionally, the backend system had inaccuracies in calculating user Agent quotas, resulting in some users having their privileges mistakenly deducted. To address these issues, Moonshot decided to reset all users' monthly quotas back to 100% and clear the used amount, allowing users to quickly return to a smooth experience.
It is reported that Kimi K2.6 has performed well in multiple tests, including the "Ultimate Human Exam" at a doctoral level and SWE-Bench Pro, which assesses real software engineering capabilities, with results comparable to or better than current leading closed-source models such as GPT-5.4, Claude Opus4.6, and Gemini3.1Pro. Since its release, the features of Kimi K2.6 have attracted a large number of users, and all users can currently access it through the official website, the latest version of the Kimi app, and Kimi API and Kimi Code programming assistant.
Moonshot stated that they will continue to monitor user feedback and strive to provide a better service experience to meet growing demands.
Key Points:
🌟 After the launch of Kimi K2.6, a surge in user traffic led to functional abnormalities and mistaken deductions of user rights.
🔧 Moonshot decided to reset quotas for all users, restoring all users' monthly quotas to 100%.
🏆 Kimi K2.6 has shown excellent performance in multiple industry tests, matching or surpassing leading models.
