The internal status endpoint (used by the IDE extensions and the API get_init_params call) now includes a model_scoped array inside the rate_limits object when the server emits per-model windows. Each entry has:
display_name— server-supplied label for the model bucket (e.g. "Fable")utilization— fraction of the weekly window consumed, or nullresets_at— ISO-8601 timestamp of window reset, or null
This field is additive: it appears only when the server returns per-model data and the list is non-empty. It gives integrators visibility into Fable 5 usage separate from the aggregate weekly limit.
Evidence
new schema field (search for "Per-model weekly windows from the server limits[] array" or "model_scoped projection failed:")
Strings lifted out of the shipped bundle, so the claim above can be checked against them.