November 2026: 15 services · five detectors · updated scoring

November 2026 · 750 outputs · 15 services

ZeroGPT detector benchmark

Model: Vendor-managed; no public version identifier. Weight in the humanization composite: 5%.

Unqualified Human Written or Most Likely Human written feedback passes. Mixed, uncertain and human labels qualified by possible AI fail. Human score is 100 minus fakePercentage, a text-share estimate. It is not GPTZero.

November 2026 · methodology v2 · provisional automated quality. 15 humanizers, 50 identical inputs each, five detectors and eight quality metrics. Read the scoring rules and post-run changes.

Small complementary signal; independent studies raise threshold and false-positive concerns. Account access does not establish calibration. Reference. None of these weights is a measured current detector accuracy.

ZeroGPT: ranked by strict-human pass rate, with raw averages separate.
HumanizerPasses/50Pass rateHuman-score average
Phrasly Ultra50/50100.00%99.10
WriteHuman46/5092.00%93.78
ZeroGPT Humanizer40/5080.00%90.75
Undetectable.ai39/5078.00%90.44
StealthGPT Super34/5068.00%83.24
Grubby22/5044.00%70.99
Rephrasy20/5040.00%66.05
BypassGPT20/5040.00%66.88
Humbot19/5038.00%70.71
Clever Humanizer18/5036.00%51.37
Solvely9/5018.00%38.09
RewriteAI7/5014.00%34.44
Grammarly Pro1/502.00%10.03
QuillBot0/500.00%2.95
HIX Bypass0/500.00%12.14

Failed strict bypass

100 minus the pass rate on these humanized outputs; this is not a false-positive rate on human writing.

ZeroGPT: failed strict bypass by service
HumanizerFailed/50Failed rate
Phrasly Ultra0/500.00%
StealthGPT Super16/5032.00%
WriteHuman4/508.00%
Rephrasy30/5060.00%
Undetectable.ai11/5022.00%
Clever Humanizer32/5064.00%
ZeroGPT Humanizer10/5020.00%
BypassGPT30/5060.00%
Grubby28/5056.00%
Grammarly Pro49/5098.00%
Humbot31/5062.00%
Solvely41/5082.00%
RewriteAI43/5086.00%
QuillBot50/50100.00%
HIX Bypass50/50100.00%

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