Hello Hugging Face Inference Providers team,
We operate a governance evaluation using sanitized, non-sensitive prompts. A provider-pinned routed call to openai/gpt-oss-120b:fireworks-ai completes and reports provider/model/usage, but the response does not expose a value that binds that response to the exact immutable revision or artifact digest served by the underlying provider.
Could you confirm whether there is a currently supported non-sensitive response field or header for:
- a request or trace identifier;
- the exact immutable served model revision or artifact digest; and
- correlation between that response and the served revision?
If not, is there a supported Inference Providers configuration that provides provider-attested served revision metadata for every call? A Hub commit alone is insufficient for this purpose because a routed provider may serve different weights.
We do not need account-specific help and will not include credentials, client content, account identifiers, or request IDs.
Thank you.