# `mix pixir.bench.subagents`
[🔗](https://github.com/Ranvier-Technologies/pixir/blob/main/lib/mix/tasks/pixir.bench.subagents.ex#L1)

Runs a deterministic Subagents benchmark and writes evidence artifacts.

This is the executable adapter for `docs/benchmarks/subagents.md`'s Pixir-native
stress layer. It intentionally avoids real provider calls by injecting fake
providers through Pixir's existing test seam.

Usage:

    mix pixir.bench.subagents
    mix pixir.bench.subagents --n 1,5,10 --repetitions 3
    mix pixir.bench.subagents --dry-run --json
    mix pixir.bench.subagents --output .pixir/benchmarks/subagents/custom-run

Artifacts:

    runs.jsonl
    summary.json
    report.md

---

*Consult [api-reference.md](api-reference.md) for complete listing*