Gets aggregated per-model token usage for a specific agent revision or all revisions.

Returns one row per `(provider, model)` pair seen in this revision's executions, summing input tokens, output tokens, and call counts so callers can rank model spend and detect drift after a revision change. Pass `*` as the revision to aggregate across every revision the agent has ever run, which is useful for top-of-funnel dashboards; pass a specific revision number to attribute usage to one deployed version. Pair with the revision cost endpoints for dollar-cost breakdowns rather than raw tokens. Without `from` and `to` the rollup covers all time; with either it covers that window, read on the same timestamps as the cost timeseries (a missing `from` is 30 days before `to`), so the two reconcile over the same range. Days kept only as daily summaries come back as `summarized` rows with no revision (all-revisions reads only).

curl -X GET "https://api.fruxon.com/v1/tenants/string/agents/string/revisions/string/modelUsages"

[
  {
    "revision": 0,
    "modelId": "string",
    "providerId": "string",
    "inputTokens": 0,
    "outputTokens": 0,
    "cachedTokens": 0,
    "cacheWriteTokens": 0,
    "thinkingTokens": 0,
    "webSearchCalls": 0,
    "executionCount": 0,
    "inputCost": 0,
    "outputCost": 0,
    "webSearchCost": 0,
    "summarized": false,
    "totalCost": 0
  }
]

{
  "type": "string",
  "title": "string",
  "status": 0,
  "detail": "string",
  "instance": "string",
  "property1": null,
  "property2": null
}