The Pulse
Moonshot Restores 25% of Kimi Quotas After K3.1 Pro Problems
Moonshot AI says subscribers on every plan will receive a one-time reset restoring 25% of their monthly Kimi quota after rate limits and unnecessary consumption on K3.1 Pro. The company also says it has adjusted plan parameters to reduce un
AI.info Team · Written with AI · fact-checked ·
“These were our issues, and we should have fixed them sooner.”
Moonshot AI, in a post on its @Kimi_Moonshot account
Moonshot AI is restoring part of subscribers’ monthly Kimi quota after acknowledging rate limits and unnecessary consumption on its K3.1 Pro subscription plan. In a post dated October 25, the company said every subscriber on every plan would receive a one-time usage reset equal to 25% of their monthly quota. It described that amount as about one week of use. The announcement followed problems tied to K3.1 Pro’s launch, according to CNA’s Reuters-sourced report.
What the 25% reset gives subscribers
The adjustment is calculated against the monthly quota, not described as a new weekly allowance. For subscribers whose recorded usage is below 25% of that monthly amount, Kimi said the usage figure would reset to zero. That distinction matters: the company announced a one-time correction to the usage meter, not a recurring weekly credit or a full monthly reset for every account.
Kimi said subscribers would see the adjustment in their usage bars by the end of the day, Pacific time. The post did not give a more precise completion time or confirm that the change had reached individual accounts. Its wording also leaves the underlying subscription quotas unchanged: the promised adjustment restores a portion of the existing monthly allowance.
Moonshot points to subagent calls and output length
The company paired the reset with changes to the plan’s parameters. It said it was reducing unnecessary subagent calls and output length so subscribers’ quotas would go further. Subagent calls are delegated work handled by other AI agents; output length refers to the amount of generated text. Both can affect how much usage a request consumes, though Moonshot did not publish before-and-after figures for either change.
The announcement does not identify which settings changed, quantify how much quota those changes might save, or specify how much of the earlier consumption came from subagent activity versus lengthy outputs. It also does not provide a detailed account of what caused the rate limits. For subscribers, the reset offers immediate usage relief; the parameter changes are the company’s stated response to future consumption.
Acknowledgment without a usage breakdown
Moonshot’s apology directly accepts responsibility for the rate limits and unnecessary quota consumption, saying the issues were its own and should have been fixed sooner. The company did not say how many subscribers were affected or publish usage data showing the scale of the problem. Those omissions make it difficult to judge the size of the correction beyond the company’s stated 25% figure.
The clearest measure in the announcement is therefore the reset itself: a one-time restoration of a quarter of monthly quota, with accounts below that usage level moved back to zero. Whether the changes to subagent calls and output length prevent similar quota pressure is not quantified in the post. Moonshot’s stated deadline for the usage bars to reflect the reset was the end of October 25, Pacific time.