Moonshot AI Apologizes for Kimi K3.1 Pro Rate Limits and Grants All Subscribers a One Time 25 Percent Monthly Quota Reset
We're sorry for the rate limits and unnecessary quota consumption on the K3.1 Pro subscription plan. These were our issues, and we should have fixed them sooner. Today, every subscriber on every plan gets a one-time usage reset, restoring 25% of their monthly quota (usage below 25% resets to zero), equal to one week of usage. Your usage bar will reflect this adjustment by end of day PST. We've also adjusted the plan's parameters and are reducing unnecessary subagent calls and output length, so your quota goes further. Thank you for your patience while we put this right.
Context
Agentic models that spawn subagents and generate long reasoning traces can burn through user subscription quotas far faster than standard chat, forcing labs to tune harness efficiency and compensate paying users when limits misfire.
Moonshot AI recently launched Kimi K3.1 Pro, a frontier reasoning and agent model that entered the top tier of the Agent Arena leaderboard, driving a surge in developer signups alongside complaints that multi step agent tasks were exhausting monthly subscription allowances rapidly.
Deep dive
Moonshot AI issued a public apology on its official Kimi X account on Sunday afternoon addressing strict rate limits and excessive quota drain experienced by paying users on the Kimi K3.1 Pro subscription plan. The company acknowledged that internal configuration issues caused the agent harness to consume far more of a subscriber's allowance than intended.
To compensate affected users, Moonshot announced a one time usage credit across every subscription tier that restores 25 percent of each subscriber's monthly quota, equivalent to roughly one full week of usage, with accounts currently under 25 percent utilization resetting all the way to zero. The company stated that subscriber usage bars would reflect the credit by the end of the day Pacific time.
Alongside the quota refund, Moonshot engineers updated K3.1 Pro's runtime parameters to curb redundant subagent spawning and trim overly verbose intermediate outputs so that multi step coding and research sessions consume fewer tokens per task.
What to expect next
Developers using Kimi K3.1 Pro will watch whether the reduced subagent call frequency preserves the model's strong Agent Arena performance while making monthly subscription allowances last through heavy daily workloads. The episode also sets a visible precedent for how frontier model providers handle token heavy agent harnesses, where a minor change in default subagent recursion can multiply user quota burn overnight unless provider side guardrails are tuned carefully.