Skip to main content

​
Agent Metrics

​
Overview

Your Agent keeps a single usage object that accumulates metrics over every run (do, print_do, stream, etc.). Use it to monitor total tokens, requests, tool calls, cost, and duration for that agent. When printing is enabled, an Agent Metrics panel is shown after each task so you can see the updated totals.

​
What You Can Access

On any Agent instance, use agent.usage. It is a RunUsage object (or None before the first run). After runs, you get:
MetricMeaning
input_tokensTotal prompt tokens across all runs
output_tokensTotal completion tokens across all runs
requestsNumber of LLM requests
tool_callsNumber of tool calls executed
costEstimated total cost (when available)
durationTotal execution time (seconds)
Other fields (e.g. cache_read_tokens, reasoning_tokens, time_to_first_token) are available on the same object when the provider reports them.

​
Example

from upsonic import Agent, Task

agent = Agent("anthropic/claude-sonnet-4-5", print=True)
agent.print_do(Task("What is 2 + 2? Reply with one number."))
agent.print_do(Task("What is 3 + 3? Reply with one number."))

# Totals for this agent across all runs
u = agent.usage
print(f"Requests: {u.requests}, Input: {u.input_tokens}, Output: {u.output_tokens}, Tool calls: {u.tool_calls}, Cost: ${u.cost:.4f}")