{"id":"humanlike-public-benchmark-summary-2026-04-20","title":"HumanLike Public Benchmark Summary Dataset","updatedAt":"2026-04-20","lastVerified":"2026-04-20","version":"2026.04.20","releaseDate":"2026-04-20","whatChanged":"Refreshed the public detector summary snapshot, aligned benchmark wording with methodology caveats, and kept JSON/CSV downloads in sync with the current public rows.","averageBypassRate":"99.2%","detectorsCovered":5,"datasetScope":"This appendix documents the current public benchmark summary rows rather than raw customer documents or per-document detector exports.","verificationScope":"Current public benchmark summary, detector-level result rows, downloadable benchmark dataset, and methodology references.","benchmarkUrl":"https://humanlike.pro/research/humanizer-benchmark-2026","methodologyUrl":"https://humanlike.pro/research/methodology","appendixUrl":"https://humanlike.pro/research/benchmark-data","downloads":{"json":"https://humanlike.pro/research/benchmark-data.json","csv":"https://humanlike.pro/research/benchmark-data.csv"},"rows":[{"detector":"GPTZero","slug":"gptzero-results","resultLabel":"Benchmark-backed result","resultValue":"99.4%","summary":"HumanLike benchmark coverage includes GPTZero because it is one of the most cited academic and creator-facing detector brands.","note":"Refer to the methodology page for sample handling, scoring definitions, and limitations.","referencePage":"https://humanlike.pro/research/gptzero-results"},{"detector":"Turnitin","slug":"turnitin-results","resultLabel":"Benchmark-backed result","resultValue":"99.2%","summary":"HumanLike benchmark coverage includes Turnitin because it is a core reference point for academic AI detection workflows.","note":"Turnitin environments vary by institution. The methodology page documents how HumanLike treats platform variability and interpretation caveats.","referencePage":"https://humanlike.pro/research/turnitin-results"},{"detector":"Originality.ai","slug":"originality-results","resultLabel":"Benchmark-backed result","resultValue":"99.0%","summary":"HumanLike benchmark coverage includes Originality.ai because it is widely referenced in publisher and SEO workflows.","note":"Originality.ai updates frequently. Use the benchmark and methodology pages together when citing results.","referencePage":"https://humanlike.pro/research/originality-results"},{"detector":"Copyleaks","slug":"copyleaks-results","resultLabel":"Benchmark-backed result","resultValue":"99.3%","summary":"HumanLike benchmark coverage includes Copyleaks because it is frequently used in institutional and enterprise review stacks.","note":"Detector behavior varies with document type and length. See methodology for inclusion rules and constraints.","referencePage":"https://humanlike.pro/research/copyleaks-results"},{"detector":"ZeroGPT","slug":"zerogpt-results","resultLabel":"Benchmark-backed result","resultValue":"99.5%","summary":"HumanLike benchmark coverage includes ZeroGPT because it remains a common public AI detector reference for non-technical users.","note":"Public-facing detector tools can change quickly. The benchmark page records the current benchmark snapshot date.","referencePage":"https://humanlike.pro/research/zerogpt-results"}]}