Skip to main content

Documentation Index

Fetch the complete documentation index at: https://concentrate.ai/docs/llms.txt

Use this file to discover all available pages before exploring further.

The model usage report is a periodic summary that breaks down your API spend by model and provider. It shows which models are costing the most, how many requests each handles, token counts, and cache hit rates. This helps you identify optimization opportunities and track spending patterns across your model portfolio.

How it works

Each evaluation cycle:
  1. Queries yesterday’s billing data from billing_daily for the target entity
  2. Groups by model and provider — aggregates total credits, requests, input tokens, output tokens, and cached tokens per model
  3. Sorts by cost descending — the most expensive models appear first
  4. Calculates total spend across all models
  5. Sends the report if there was any usage (at least one request with non-zero spend)
This is a report-style alert, not a threshold-based alert. It fires whenever there is data to report.

Configuration

You can configure the model usage report from the Alerts page in your dashboard.
SettingDescriptionOptions
EnabledToggle the report on or offtrue / false
ScopeWhat entities to report onorganization, user

Scope

  • Organization — reports on model usage across your entire organization.
  • User — reports on model usage across your personal API keys.

Report contents

The report includes a per-model breakdown with:
FieldDescription
ProviderThe AI provider (e.g., OpenAI, Anthropic)
ModelThe specific model used (e.g., gpt-4, claude-sonnet-4-20250514)
CostTotal spend for this model in dollars
RequestsNumber of API requests
Input TokensTotal input tokens consumed
Output TokensTotal output tokens generated
Cache RatePercentage of tokens served from cache

Cooldown

The model usage report has a 7-day cooldown (604,800 seconds). After a report is sent, no duplicate report is sent for the same entity for 7 days. This makes it effectively a weekly report.

Notifications

  • Email — includes a styled table with the full model breakdown, total spend, and a link to your usage details page
  • SMS — a shorter summary listing total spend and the top 3 models by cost