Review Benchmarks — reproducible checks and AI search observationsSkip to content (#main)Review BenchmarksMethods, controls & recorded results (/)Research index (/)Code review evaluation (/evaluations/order-service/)Citation study (/research/citations/)

Methods, controls & recorded results

Inspect the method behind the result.

Pinned code changes, disclosed requirements, clean controls and recorded outputs. Read what was tested and which evaluations are still pending.

Inspect the evaluationRequirements, public changes and captured findings. (/evaluations/order-service/)Trace the citationsBuyer queries, observed answers and cited source URLs. (/research/citations/)

Matched public behavior requirements (/evaluations/order-service/)

Four disclosed requirements, immutable public changes, executable controls and captured Perch results. CodeRabbit and Greptile reviews remain unscored until authenticated runs complete.

AI-search citation baseline (/research/citations/)

Nine buyer queries, twelve observed answers, source URLs and the gaps found in the reviewed documents. This baseline predates the comparison rewrite.

Published by Perch / Verglas LLC.

Review BenchmarksMethods, controls & recorded results.

Published by Perch · © 2026 Verglas LLC. Review Benchmarks is a Perch publication.

Perch’s MIT license covers the CLI; hosted model analysis is a paid service. Relevant source is sent to the configured model service.

Perch CLI MIT license (https://github.com/lakeday-org/perch/blob/main/LICENSE) · Article text (https://reviewbenchmarks.dev/index.md) · Resource index (/llms.txt)