Open evidence

AI support benchmarks

Versioned technical tests with downloadable inputs, machine-readable results, explicit limitations, and no claim that synthetic performance equals customer value.

Published 2026-07-24

Qwen3 support analysis benchmark on 60 synthetic messages

Three independent ReplyPilot runs measuring topic stability, risk output, generated assets, latency, token usage, and estimated inference cost.

View benchmark

Synthetic messages

3/3

Topic agreement

12.0s

Average latency