Claude CodeExperimentsBlogPortfolioAbout me

Experiments / Traces

Back to Experiments

Agent Trace Viewer

Step-by-step recordings of real agent runs: thinking, tool calls with arguments, results, latency, tokens, cost.

claude-opus-4-7success

Research agent: latest agent eval papers

Find 3 papers from 2026 on benchmarking AI agents, extract methodology, summarize.

duration
18.42s
tool calls
6
tokens
11.2k
cost
$0.0834