Rate Limits

Rate limits on Runars are tied to your organization and scale with your lifetime spending. This page covers tier thresholds, how limits work, and the headers you'll receive in response.

Rate limit tiers

Your current tier and progress toward the next tier are displayed in Dashboard → Rate Limits.

Rate limit headers

Every response (successful or rate-limited) includes rate-limit headers for the current request:

Note: Output token limits are tracked after each request completes (since output tokens aren't known until after the response is generated). The OTPM headers mirror the RPM and ITPM format:

• anthropic-ratelimit-output-tokens-limit
• anthropic-ratelimit-output-tokens-remaining
• anthropic-ratelimit-output-tokens-reset

On a 429 response, an additional retry-after header (in seconds) indicates how long to wait before retrying.

OTPM rate limit blocking

Output tokens are only known after a request completes, so the OTPM limit cannot block the request that caused an overage—it can only block subsequent requests in that minute window. When this happens:

1. Your next request after the OTPM limit is hit will receive a 429 response with the retry-after header.
2. Your organization's admins and members receive an email notification (at most once per hour) alerting them to the OTPM block.
3. Requests resume normally in the next 60-second window.

Rate limit errors

When you exceed a rate limit, the API returns a 429 status code with an error response:

Anthropic-compatible endpoints (/v1/messages, /v1/messages/count_tokens):

{
  "type": "error",
  "error": {
    "type": "rate_limit_error",
    "message": "Rate limit exceeded..."
  }
}

OpenAI-compatible endpoints (/v1/chat/completions, /v1/responses):

{
  "error": {
    "type": "insufficient_quota",
    "message": "Rate limit exceeded..."
  }
}

What counts against rate limits

• Requests: Each API call (even non-streaming requests count as one request).
• Input tokens: All tokens in your request's messages, including system prompts and conversation history.
• Output tokens: Only tokens actually returned by the model—not your max_tokens limit.

Exempt endpoints

The count_tokens endpoint (POST /v1/messages/count_tokens) does not count against rate limits. Use it freely to pre-estimate costs or validate token counts before making billable requests.

Tracking your usage

Visit Dashboard → Usage to see:

• Current credit balance
• Usage this month (tokens and cost breakdown)
• Lifetime spending (used to determine your tier)
• Billing history

For request-level details, check Dashboard → Logs.

See also

• Pricing — Credit model and how to add funds
• Errors — Full error reference including rate-limit errors