Usage limits
Per-API-key request limits and budget caps.
Rate limit#
| Target | Default limit | Notes |
|---|---|---|
| Chat Completions (per API key) | 60 RPM | Contact support for a higher limit |
When a limit is exceeded, 429 Too Many Requests is returned.
Budget caps#
You can set daily, weekly, and monthly budget caps (credits) per API key (daily_budget_krw / weekly_budget_krw / monthly_budget_krw). When usage in that window plus the estimated cost of a call would exceed the cap, 402 is returned. The daily cap is a rolling 24 hours and the weekly cap is a rolling 7 days (UTC), not a calendar week. Independently of money, you can also set a monthly token limit (combined input+output, monthly_token_limit) — exceeding it returns 402 monthly_token_quota_exceeded.
Key scope restrictions#
Per key, you can restrict the allowed model list. For production keys, we recommend allowing only the models you need — violations return
400.Request limits#
max_tokens: 1 or more (default 4096). A value above the model'smax_outputis lowered to that maximum instead of failing.400is returned only when input plus output cannot fit the context windowtemperature: 0.0 – 2.0- Context length: see each model's
context_window - Prepaid balance cap for top-ups: 2,000,000 credits by default, rising to 5,000,000 credits as your cumulative payments grow and to 20,000,000 credits after business verification