Enforced billable-query limit for native Google Search with Gemini 3.8 Flash

We use gemini-3.8-flash with native googleSearch through v1beta/models.generateContent and Python SDK google-genai==1.73.1.

We need to preserve search grounding while enforcing a cumulative financial ceiling per asynchronous job, including tokens, search charges, retries and uncertain requests.

An in-memory transport test shows that setting AutomaticFunctionCallingConfig.maximum_remote_calls to 1 or 2 produces identical REST request bodies: this client-side counter is not transmitted as a native Search query limit. We found no billable-query maximum in the documented GoogleSearch schema. The documented project spend cap has approximately ten minutes of enforcement latency.

Is there a supported, server-enforced maximum for billable Search queries within one request? Please identify the exact field, supported API/model and enforcement semantics. Alternatively, is there an adjustable quota that prevents each internally billable query, rather than merely limiting Search-enabled GenerateContent requests? Please distinguish these quota units explicitly.

We would appreciate a supported route or an explicit statement if this bound is not available. No credentials, project identifiers or customer data are included.