November 2026 · Independent humanizer benchmarks. Explore the evidence

Compare AI humanizers

Pick two to four entries and inspect them side by side. All numbers come from the same shared source set and measurement profiles. An asterisk marks detector-dependent, provisional values.

Pick 2 to 4 entries 1 selected

November 2026
Qwen 2.5 14B HumanizerVerified by HumanizeBenchWe ran this model ourselves on the benchmark source examples and checked its execution records. Verification confirms the tested model’s provenance; it does not guarantee performance or imply that every detector returned a result.arunsingh80475/qwen2.5-14b-humanizer-v4-lora
Open weights
Shared score components
Overall score43.12
Panel bypass score66.2%
Meaning preservation21.1%
Raw meaning cosine0.7892
Readability78.1%
Category consistency84.0%
Detector coverage230/231
Identical outputs0/33
Penalty points−15
Per-detector human-likelihood scores · measured outputs only
GPTZero29.3% 33/33
ZeroGPT30.1% 33/33
Copyleaks66.7% 33/33
Originality.ai23.1% 32/33
Pangram71.1% 33/33
Grammarly88.6% 33/33
QuillBot99.8% 33/33
Category scores · before entry-level penalties
Academic essay53.8
Application essay41.5
Blog post55.7
Business email64.1
Marketing copy75.9
Discussion board56.8
News article64.9
EvidenceAll 33 outputs