November 2026 · 750 outputs · 15 services
GPTZero detector benchmark
Model: 5.0b / 2026-11-04-base. Weight in the humanization composite: 30%.
predicted_class=human; mixed and ai fail. Human score is a class probability, not the percentage of human-written words.
A substantial second signal, informed by independent robustness and false-positive research; current-version shared calibration remains undone. Reference. None of these weights is a measured current detector accuracy.
| Humanizer | Passes/50 | Pass rate | Human-score average |
|---|---|---|---|
| Phrasly Ultra | 50/50 | 100.00% | 99.31 |
| WriteHuman | 40/50 | 80.00% | 75.52 |
| StealthGPT Super | 30/50 | 60.00% | 52.21 |
| Rephrasy | 22/50 | 44.00% | 38.24 |
| Undetectable.ai | 16/50 | 32.00% | 26.50 |
| ZeroGPT Humanizer | 15/50 | 30.00% | 25.45 |
| Clever Humanizer | 13/50 | 26.00% | 21.86 |
| Humbot | 5/50 | 10.00% | 8.51 |
| BypassGPT | 4/50 | 8.00% | 5.44 |
| Grubby | 4/50 | 8.00% | 6.87 |
| Grammarly Pro | 0/50 | 0.00% | 0.00 |
| Solvely | 0/50 | 0.00% | 0.00 |
| RewriteAI | 0/50 | 0.00% | 0.00 |
| QuillBot | 0/50 | 0.00% | 0.00 |
| HIX Bypass | 0/50 | 0.00% | 0.00 |
Failed strict bypass
100 minus the pass rate on these humanized outputs; this is not a false-positive rate on human writing.
| Humanizer | Failed/50 | Failed rate |
|---|---|---|
| Phrasly Ultra | 0/50 | 0.00% |
| StealthGPT Super | 20/50 | 40.00% |
| WriteHuman | 10/50 | 20.00% |
| Rephrasy | 28/50 | 56.00% |
| Undetectable.ai | 34/50 | 68.00% |
| Clever Humanizer | 37/50 | 74.00% |
| ZeroGPT Humanizer | 35/50 | 70.00% |
| BypassGPT | 46/50 | 92.00% |
| Grubby | 46/50 | 92.00% |
| Grammarly Pro | 50/50 | 100.00% |
| Humbot | 45/50 | 90.00% |
| Solvely | 50/50 | 100.00% |
| RewriteAI | 50/50 | 100.00% |
| QuillBot | 50/50 | 100.00% |
| HIX Bypass | 50/50 | 100.00% |
This measures bypass of the tested detector, not its false-positive rate or accuracy on independently verified human writing. A high pass rate alone cannot establish that the detector is bad.
Per-output scores CSV · Frozen inputs and prompts CSV · Scoring manifest · Verification bundle · GitHub report · Earlier editions