November 2026 · Independent humanizer benchmarks. Explore the evidence

Pangram detector results

Mean reported human-likelihood scores for each entry on available Pangram measurements. Higher values mean the detector assigned more human likelihood, not that authorship was established.

758/759Measurements collected
23Entries in the shared panel
33Sources per entry
PolygraphActive collection route
Rank ⓘHumanizerPangram ⓘ ↓Bypass ⓘMeaning ⓘReadability ⓘPenalties ⓘCoverageEvidence
133 examples
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
71.0866.221.178.1△ −15.0
230/231
233 examples
Rephrasyrephrasy.ai
67.2578.035.773.4△ −12.0
231/231
333 examples
CleverHumanizercleverhumanizer.ai
67.2077.325.475.8△ −10.0
231/231
433 examples
UndetectedGPTundetectedgpt.ai
60.8261.346.879.8△ −10.0
231/231
533 examples
StealthGPTstealthgpt.ai
59.0138.646.479.9△ −12.0
231/231
633 examples
Superhumanizersuperhumanizer.ai
47.9646.543.274.6△ −10.0
231/231
733 examples
HIX Bypasshixbypass.com
42.2145.441.376.9△ −20.0
231/231
833 examples
Humanizer 12B · OptiQ 4-bitVerified 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.mlx-community/humanizer-12b-OptiQ-4bit
35.7236.360.977.1△ −7.0
231/231
933 examples
Humanizer 12B · GGUFVerified 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.jialinyyzz/humanizer-GGUF
33.3225.569.680.0△ −2.0
231/231
1033 examples
Humanizer 12BVerified 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.jialinyyzz/humanizer
33.3225.569.680.0△ −2.0
231/231
1133 examples
Humanizer 12B · MLX 8-bit (cowWhySo)Verified 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.cowWhySo/humanizer-12b-mlx-8bit
31.7725.366.579.0△ −4.0
230/231
1233 examples
Humanizer 12B · MLX 8-bit (npario)Verified 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.npario/humanizer-12b-mlx-8bit
31.2424.665.978.8△ −5.0
229/231
1333 examples
GPTinfgptinf.com
30.8537.741.471.7△ −10.0
231/231
1433 examples
Humanizer · CMFVerified 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.infosave/humanizer-cmf
30.8330.265.978.0△ −3.0
231/231
1533 examples
SmartHumanizersmarthumanizer.ai
29.4929.050.277.6△ −10.0
231/231
1633 examples
Smaug 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.abacusai/Smaug-Humanizer
25.6218.576.581.3△ −1.0
231/231
1733 examples
Humanizer 1B · OptiQ 4-bitVerified 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.mlx-community/humanizer-1B-OptiQ-4bit
4.370.677.171.2△ −17.0
229/231
1833 examples
HumanizerAI · GGUFVerified 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.mradermacher/humanizerAI-GGUF
2.221.075.982.6△ −24.0
199/231
1933 examples
Rehumanizerehumanize.io
2.150.860.282.2△ −6.0
231/231
2033 examples
Undetectable AIundetectable.ai
1.510.555.679.6△ −18.0
231/231
2133 examples
GoHumanizeVerified 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.gohumanize/gohumanize-open-humanizer
0.120.292.085.1△ −7.0
231/231
2233 examples
Open Humanizer 4BVerified 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.akhaliq/open-humanizer-4b
0.100.096.685.1△ −10.0
231/231
2333 examples
Humanizer 3BVerified 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.mondk/Humanizer-3B-Safetensors
0.001.312.372.1△ −20.0
208/231

Detector values are mean reported human-likelihood scores; 100 is highest. Scores use available detector measurements. Coverage and methodology are available above.