AI Humanizer BenchmarkLeaderboard / Compare / StealthGPT vs SuperHumanizer
Head to head · October 2026 cycle
StealthGPT vs SuperHumanizer
On the October 2026 leaderboard, SuperHumanizer leads StealthGPT 78.23 to 64.52, a gap of 13.71 points. SuperHumanizer ranks #6 and StealthGPT ranks #9 out of 11 tools. SuperHumanizer scores higher on all 7 detectors and in all 7 writing categories. StealthGPT is ahead on readability (80.6 vs 71.0). On bypass rate, StealthGPT scores 62.8% and SuperHumanizer scores 83.9%, but their 95% confidence intervals overlap, so that difference is not statistically clear.
- Detectors
- 0 – 7
- Categories
- 0 – 7
- Components
- 1 – 4
- Last tested
- Prompts
- 33
- Methodology
- v1.1.0
Score components
The overall score weights bypass rate at 42%, meaning preservation at 32%, readability at 16% and consistency across writing categories at 10%. Penalties for output-quality failures are then subtracted. Full scoring formula
Overlapping CIs means the two tools' 95% confidence intervals for bypass rate overlap, so the bypass difference is not statistically clear.
Detector by detector
Each figure is one detector's average human-likelihood score across all of a tool's outputs, on a 0–100 scale. Higher means more human. SuperHumanizer scores higher on all 7 detectors.
Writing categories
Each category score applies the overall-score weights to that category's prompts only. Bypass, meaning preservation and readability are averaged within the category, and the tool's overall consistency score is added. Penalties are not subtracted. Categories with few prompts give less reliable scores, so each row shows its prompt count.
Where each one wins
Measures where one tool is ahead of the other. Ties are left out. Score components come first, then detectors, then categories, with the largest gap first in each group.
StealthGPT
1 win- Readability Score80.6 vs 71.0
SuperHumanizer
18 wins- Bypass rate Score83.9 vs 62.8
- Consistency Score85.6 vs 75.9
- Meaning preservation Score75.2 vs 67.7
- Penalties Score−1.0 vs −4.0
- Originality.ai Detector88.2 vs 62.7
- QuillBot Detector81.3 vs 61.3
- ZeroGPT Detector83.5 vs 67.4
- Grammarly Detector83.5 vs 69.0
- Winston AI Detector68.8 vs 55.3
- GPTZero Detector47.6 vs 41.5
- Copyleaks Detector63.6 vs 60.6
- Application essay Category76.8 vs 51.2
- Discussion board Category85.7 vs 69.6
- News article Category85.3 vs 70.8
- Business email Category66.6 vs 59.5
- Blog post Category76.9 vs 70.1
- Academic essay Category82.6 vs 78.3
- Marketing copy Category80.6 vs 77.3
Questions people ask
Is StealthGPT better than SuperHumanizer?
On the October 2026 leaderboard, SuperHumanizer has an overall score of 78.23 and StealthGPT has 64.52, a gap of 13.71 points. SuperHumanizer ranks #6 and StealthGPT ranks #9 out of 11 tools. SuperHumanizer scores higher on all 7 detectors and in all 7 writing categories. Both tools rewrote the same 33 prompts, the same 7 detectors checked every output, and scoring followed methodology v1.1.0. All inputs, outputs and detector verdicts are public.
Which bypasses AI detectors better, StealthGPT or SuperHumanizer?
SuperHumanizer has the higher bypass rate, 83.9% vs 62.8% for StealthGPT. It scores higher on all 7 detectors. The two tools' 95% confidence intervals for bypass rate overlap, so the difference is not statistically clear.
Do StealthGPT and SuperHumanizer keep the original meaning?
Meaning preservation measures how close each output stays to its input, using embedding similarity on a 0–100 scale. SuperHumanizer scores 75.2 and StealthGPT scores 67.7. On readability, StealthGPT scores higher, 80.6 vs 71.0. This cycle, StealthGPT lost 4.0 points to penalties and SuperHumanizer lost 1.0 points to penalties.