Files
Assistant/docs/MODEL_FLEET_POLICY.md
m 30443027a6
CI / test (push) Failing after 37s
CI / test (pull_request) Failing after 32s
feat: add evidence-based adaptive model rating
2026-08-22 23:42:10 +03:00

1.2 KiB

Model fleet policy

Domovoy does not equate intelligence with GPU, model name, or provider kind. Hardware and endpoint identity are operational observations only. A model is eligible for routing only after the existing qualification and activation gates pass.

FAST, MEDIUM, and HEAVY are capability service classes, not hardware claims. A CPU model may be the best available quality model, and a GPU model may be unavailable, degraded, or unrated.

The Capability owner derives a provider-neutral ModelRating from evidence:

  • QUALITY: high measured quality and zero critical failures;
  • BALANCED: supported model with usable quality/reliability evidence;
  • FAST: supported model whose observed latency is strong enough for quick work;
  • UNRATED: discovery or incomplete evidence; it cannot be promoted by name.

The score combines quality, reliability, latency and evidence coverage. Missing measurements remain explicit. A model can change category after a new immutable qualification/performance revision. The current CPU route is an operational choice, not a permanent hardware assumption. CPU is preferred in the current deployment; if it is unavailable, the policy tries another ready class before using the Lifeboat recovery path.