POST /v1/messages/count_tokens

Request

curl https://api.model-gate.com/v1/messages/count_tokens \
  -H "x-api-key: mg_live_..." \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "ch-46s",
    "messages": [{"role": "user", "content": "Count these tokens."}]
  }'

Response

{
  "input_tokens": 9
}

Aliases are resolved exactly as for /v1/messages. This endpoint does not consume the API-key RPM counter, but authentication and model availability are still validated.

Error

{
  "error": {
    "type": "invalid_request_error",
    "code": "model_not_found",
    "message": "Unknown or disabled model: unknown-model"
  }
}