Agent quota & billing

Agent mode uses Kimi's unified quota system. All membership features (Agent, Deep Research, Slides, Docs, Sheets, Kimi Code, Kimi Claw, etc.) share a single quota pool, with quota deducted based on actual token consumption.

Billing model

  • Unified quota: All membership features share one quota pool — allocate usage however you like
  • Pay-per-use: Quota consumption depends on task complexity and duration (i.e., token usage) — simple tasks cost less, complex tasks cost more
  • Monthly refresh: Quota resets monthly, aligned with your subscription cycle
  • Usage priority: Bonus quota (e.g., trial credits, promotional rewards) is consumed first, followed by plan quota

Example: With a Moderato plan, generating a simple PPT might consume about 1–2% of quota, while a single Deep Research session might use about 5–10%.

What happens when quota runs out?

When your quota is exhausted:

  • Any task currently in progress will complete normally
  • New tasks will show an "insufficient quota" notice
  • Your options:
    • Wait for your monthly quota to auto-refresh
    • Upgrade to a higher-tier membership for more quota

How to check quota usage?

  • Web: Profile → Settings → Subscription
  • App: Profile → Membership Plan → Subscription What you can view:
  1. Current quota balance (as a percentage)
  2. Next refresh date
  3. Last 10 usage records (timestamp, feature used, consumption percentage)

Usage records may have a brief delay. Refer to the actual quota display for the most current information.

Does a failed task still consume quota?

  • Quota is deducted after a task executes successfully, based on actual consumption
  • If a task fails due to a system error (no valid result returned), click the 👎 button to report it. After verification, the corresponding quota will be refunded
Agent quota & billing - Kimi Help Center