Fix OpenRouter Rate Limit & "model not allowed" Errors

OpenClaw with OpenRouter can fail with two related errors: model not allowed (the model requires higher account access) or 429 Too Many Requests (you hit the rate limit). This guide shows how to diagnose and fix both.

Identify the Error Type

Fix: model not allowed (403)

Check your OpenRouter account at openrouter.ai/account and verify:

1. Credit balance — Most models require a positive credit balance. Add at least $5 to unlock pay-as-you-go models.
2. Model availability — Some models (like Claude Opus) require a minimum balance or are restricted to certain regions.
3. API key scope — Regenerate your API key if it was created under a different account tier.

As a quick fix, switch to a model you know you have access to:

# openclaw.json — switch to a model available on free tier
{
  "model": "openrouter/auto",
  "parameters": {}
}

# Or use a specific free model:
{
  "model": "mistralai/mistral-7b-instruct",
  "parameters": {}
}

Fix: Rate Limit Exceeded (429)

If you're on the free tier, the limit is 10 requests/minute. To increase it:

Configure Model Fallbacks

OpenClaw supports automatic model fallback. When the primary model returns a rate-limit or access error, it will try the next model in the list:

{
  "model": "anthropic/claude-sonnet-4-6",
  "fallbackModels": [
    "openai/gpt-4o",
    "google/gemini-2.0-flash",
    "openrouter/auto"
  ]
}

Tune Request Rate

{
  "model": "openrouter/auto",
  "advanced": {
    "maxConcurrentAgents": 1,
    "requestDelay": 1500
  }
}