# Track cost, tokens, and latency per model and per component

> The metrics system has been redesigned to provide granular, per-model and per-component tracking across the entire agent, team, and workflow lifecycle.

- Published: 2026-02-24
- Author: Agno Team
- Category: Changelog
- Canonical: https://agno-com-nine.vercel.app/articles/track-cost-tokens-and-latency-per-model-and-per-component
- Markdown: https://agno-com-nine.vercel.app/articles/track-cost-tokens-and-latency-per-model-and-per-component.md

The metrics system has been redesigned to provide granular, per-model and per-component tracking across the entire agent, team, and workflow lifecycle. Instead of a single aggregate view, you now get a clear breakdown of where resources are consumed — improving cost attribution, capacity planning, and performance optimization.

Details**:**

- Per-model metrics for accurate cost allocation and provider comparison
- Per-component tracking across agents, teams, and workflows for end-to-end visibility
- Consistent data across sync, async, and streaming execution paths

View the [metrics docs](https://docs.agno.com/reference/run/metrics) for more.