# Benchmark evidence

The benchmark is a deterministic smoke benchmark for the production quote
engine, not a synthetic microbenchmark. It processes the checked-in three-item
manifest 1,000 times, for 3,000 quote operations, and prints a checksum so a
run cannot silently skip work.

## Reproduction

```text
moon --version
moon run benchmarks
```

Expected checksum:

```text
items=3000 total=CNY 269930.00
```

Validation record from 2026-08-24: Windows host, MoonBit `0.1.20260819` /
moonc `v0.10.9+6e6c44045`, wasm-gc target. Three command wall times were
194.29 ms, 177.00 ms, and 163.36 ms; median 177.00 ms. This is 3,000 quote
operations, approximately 16,949 operations per second. The first run
includes process or build warm-up, so the median is the reported comparison
value. Re-run the command on the target host before making performance claims.