Open-weight AI humanizer benchmark
Compare 13 open-weight entries evaluated on the same 33 source texts. The table ranks saved outputs by the November 2026 composite, which includes detector scores, meaning preservation, readability, consistency, and penalties.
Humanizer · CMF leads the open-weight entries in this snapshot with a composite of 45.23 out of 100. Related model formats are listed individually because each produced its own saved outputs.
| Rank ⓘ | Humanizer | Overall ⓘ ↓ | Bypass ⓘ | Meaning ⓘ | Readability ⓘ | Penalties ⓘ | Coverage | Evidence |
|---|---|---|---|---|---|---|---|---|
| 133 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 | 45.23 | 30.2 | 65.9 | 78.0 | △ −3.0 | 231/231 | |
| 233 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 | 45.19 | 25.5 | 69.6 | 80.0 | △ −2.0 | 231/231 | |
| 333 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 | 45.19 | 25.5 | 69.6 | 80.0 | △ −2.0 | 231/231 | |
| 433 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 | 44.18 | 18.5 | 76.5 | 81.3 | △ −1.0 | 231/231 | |
| 533 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 | 43.13 | 36.3 | 60.9 | 77.1 | △ −7.0 | 231/231 | |
| 633 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 | 43.12 | 66.2 | 21.1 | 78.1 | △ −15.0 | 230/231 | |
| 733 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 | 41.93 | 25.3 | 66.5 | 79.0 | △ −4.0 | 230/231 | |
| 833 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 | 39.76 | 23.7 | 65.9 | 78.8 | △ −5.0 | 231/231 | |
| 933 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 | 32.90 | 0.2 | 92.0 | 85.1 | △ −7.0 | 231/231 | |
| 1033 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 | 30.91 | 0.0 | 96.6 | 85.1 | △ −10.0 | 231/231 | |
| 1133 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 | 17.87 | 0.8 | 77.1 | 71.2 | △ −17.0 | 231/231 | |
| 1233 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 | 11.71 | 0.2 | 75.9 | 82.6 | △ −24.0 | 229/231 | |
| 1333 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 | 0.00 | 0.1 | 12.3 | 72.1 | △ −20.0 | 231/231 |
Gold, silver and bronze highlight the current top three overall values. All values use our measurements. Scores use available detector measurements. Coverage and methodology are available above.
No entries match these filters.
Explore the results
Compare the open-weight and commercial cohorts or download the November dataset.