HumanLike benchmark data

This page exposes the public benchmark summary dataset used on the HumanLike research pages. Use it when you need the current detector-level result rows, downloadable JSON or CSV files, and one-step access to the methodology page that explains how to interpret the benchmark summary.

If you only need the current benchmark summary, use Humanizer Benchmark 2026. If you need the rules behind the numbers, use Research Methodology.

• Download JSON: machine-readable benchmark summary rows.
• Download CSV: spreadsheet-friendly benchmark export.
• Methodology: how the benchmark is defined and how to read the data safely.

• Use the benchmark page for the overall average benchmark result.
• Use this appendix when you need downloadable detector-level summary rows.
• Use the methodology page before treating any benchmark percentage as a universal guarantee.
• Use pricing and brand facts pages for plan limits, language support, and product facts.

The appendix is meant for inspection, not hype. If a number matters, check the corresponding reference page before quoting it elsewhere.