Lets a fast executor model ask a higher-intelligence advisor model for guidance while it works. Each advisor is any OpenRouter model, can carry its own instructions, and remembers prior consultations when the transcript is replayed.
For decision points: before committing to an approach, when your model is stuck, or before it declares a task done.
1
Your model asks the advisor for guidance, with a prompt.
2
The advisor model thinks it through and returns its advice as the tool result.
3
Your model continues with the advice and writes the final answer.
Providers and pricing
OpenRouter
Model token usage
Billed as normal token usage on each advisor model consulted.
Standard model token costs apply to every request.
Using this tool
OpenRouter runs the tool during the request and returns the result to the model, so no client-side tool loop is needed.
Supported APIs: Chat Completions, Responses, and Anthropic Messages.
Anthropic Messages Use the native Anthropic shape, { "type": "advisor_20260301", "name": "advisor", "model": "<advisor model>" }. The model is the only setting it carries, so use openrouter:advisor on Chat Completions or Responses for instructions and the other parameters.
shell
curl https://openrouter.ai/api/v1/chat/completions \
-H "Authorization: Bearer $OPENROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "openai/gpt-5-mini",
"tools": [
{
"type": "openrouter:advisor",
"parameters": {
"model": "~anthropic/claude-opus-latest"
}
}
],
"messages": [
{
"role": "user",
"content": "Design a rate limiter for a multi-tenant API. Ask your advisor if you get stuck."
}
]
}'