Rate Limits

Limits are enforced per user, not per API key. If you create 20 keys, all 20 share the same quota below.

These limits only apply to :free models. Any paid model (paid: true) has no request cap — you're billed per token from your wallet until the balance runs out. See the Free Plan page for the full breakdown of what's free.

10,000 credits/day isn't a flat token count — each free model has its own credit cost per 1,000 tokens (input + output combined), so how many tokens that buys you depends on which free model you use. See Free Plan for the per-model rates.

Per-model token limit

On top of the limits above, every :free model has its own tokens-per-minute (TPM) ceiling, input and output combined. This cap is enforced separately per model, so burning through the TPM limit on one free model doesn't touch your quota on another. All 12 free models currently share the same 12,000 tokens/minute ceiling — there's no per-model variation here, regardless of how much each model costs in daily credits (see the credit-rate table below):

The "credits / 1k tokens" column is what determines how far your shared 10,000 daily credits go — a cheaper model like mistral-nemo:free (100 credits/1k) stretches to roughly 100,000 tokens/day, while a pricier one like deepseek-v4-flash-0731:free or gpt-oss-120b:free (600 credits/1k) only gets you around 16,600 tokens/day at the same credit budget. The 12,000 TPM ceiling, on the other hand, is flat across all of them — it caps burst speed within a single minute, not your total daily usage.

What happens when you hit a limit

The request is rejected with 429, type: "rate_limit_error", and a Retry-After header (in seconds) telling you when to try again:

{
  "error": {
    "message": "You have reached your daily token limit (100,000 tokens/day).",
    "type": "rate_limit_error",
    "param": null,
    "code": null
  }
}

Messages by limit reason

• Requests/minute — Rate limit reached for requests. Limit: 10/min.
• Daily credits — You have reached your daily credit limit (10,000 credits/day — this model costs N credits per 1k tokens, input + output combined).
• Shared daily requests (Free-plan accounts) — Daily usage limit reached for free models (shared with the web chat — 50/day on the free plan). Try again tomorrow or upgrade to Pro.
• Tokens/minute (per model) — Rate limit reached for model '<model>'. Limit: 12,000 tokens/min (input + output). Applies to any of the 12 :free models, not just one — the ceiling is the same 12,000 tokens/min across all of them.

Checking your remaining quota

Rate limits aren't currently returned as response headers on /v1/* — check your current usage under Developers → API on dravenai.lat, which shows requests and credits used today plus a usage trend.