Retrieves AI model efficiency and cost metrics for the authenticated user's organization.

Returns paginated model-level metrics

includePopulationAggregates: when true, the envelope adds mostEfficientModel,
the model with the lowest positive cost per 1M tokens among every model the filters
match (the set page.totalElements counts), not only this page. Attribution buckets
such as __organization_discount__ are never chosen. Omitted when false (default).

sort: totalCost (default, descending), provider, model, usageType,
percentage, requestCount, totalTokens, uncachedInputTokens,
cachedInputTokens, cacheWriteTokens, outputTokens, imageCount, costPer1M,
avgTokensPerRequest, creditsApplied. These are the aggregate names, not the
response field names, and an unknown key fails the request with 400.

All costs are in USD. Token counts include all token types (input, output, cache).