Protocol

Evaluate the same agent task across multiple trials

A nondeterministic system needs more than one audition.

When it fits

  • One successful run is being treated as proof that an agent workflow is reliable.

When to avoid it

  • Trial count should match cost and risk; small samples still leave uncertainty.

Why it matters

Run representative tasks multiple times and report distribution, not only pass@1 from a favorable run. Capture success, failure mode, steps, cost and variance. This distinguishes occasional capability from repeatable reliability, especially on longer tasks.

Steps

  1. Reliability claims include repeated trials and failure distribution rather than one anecdote.

An example

Run the same repository change ten times in isolated copies to see whether the agent consistently edits the right files and completes tests.

Check your result

Reliability claims include repeated trials and failure distribution rather than one anecdote.

Keep this limit in mind

  • Trial count should match cost and risk; small samples still leave uncertainty.

Connected ideas

Useful with
Keep an eval set that can embarrass the agent

Evidence and sources

Supports

Anthropic's agent-eval guidance defines separate trials because model outputs vary and recommends multiple trials for more consistent evaluation.

Trial count should match cost and risk; small samples still leave uncertainty.

Demystifying evals for AI agents · See source record

All sources (1)