Moonshot AI has apologized to subscribers after the launch of its Kimi K3.1 Pro model led to rate limiting and higher-than-expected quota use. In a message posted on its Kimi account on X, the Beijing-based company said it would reset every subscriber’s monthly quota by 25 percent, an amount it described as equal to about one week’s usage.
Subscribers who had used less than 25 percent of their quota will see their usage reset to zero, the company said. Moonshot also said it has adjusted its settings to reduce unnecessary calls between cooperating agents and to trim output length.
The apology matters because launch demand has strained Moonshot before. The company previously ran into computing capacity problems after the earlier Kimi K3 release, and Chinese AI labs are widely reported to face computing shortages linked to United States chip export controls.
For users, the reset is a goodwill gesture rather than a fix for the underlying constraint. Rate limits tend to return when demand spikes again, so subscribers who rely on Kimi for heavy coding or workflow automation should expect occasional slowdowns around major releases. Comparable open models such as DeepSeek and mainstream assistants such as ChatGPT offer alternative routes when one service is under strain, though each comes with its own limits and pricing.
The episode is a reminder of how quickly a popular model launch can outrun its infrastructure. A model’s quality means little if users cannot reach it reliably in its first days, and the companies that handle that surge gracefully are the ones users tend to stick with.
Source: Reuters via Y’all 106.7
