Skip to main content
MetrixLLM calculates the exact cost of every request based on the upstream provider’s published pricing for that specific model. These costs are aggregated in real time on the Cost Analytics dashboard.

​
Cost breakdown

Because MetrixLLM tracks all metadata, you can group and breakdown your spend by any dimension:
  • Provider: Compare total spend across OpenAI vs Anthropic vs Google.
  • Model: See the precise cost split between gpt-4o and gpt-4o-mini.
  • API Key: Attribute costs to specific client applications or internal teams based on the MetrixLLM key they used.
  • Custom Properties: Group costs by specific customers or environments using Custom Properties.

​
Cache savings

The Semantic Cache reduces your upstream provider costs. The analytics dashboard prominently displays Cache Savings, showing exactly how much money and how many tokens were saved by serving responses from the edge instead of the provider.

​
Exporting data

Enterprise customers can export cost analytics data via CSV directly from the dashboard, or use the REST API to pull aggregated metrics into internal BI tools (e.g. Looker, Tableau).