HumanizerEval

Leaderboard / Compare / WriteHuman vs Grubby

September 2026 review · same 49 texts

WriteHuman vs Grubby

In the September 2026 HumanizerEval review, WriteHuman and Grubby rewrote the same 49 English texts. WriteHuman cleared all 5 detectors on 32 of them; Grubby on 17. The difference is mostly Pangram: on 17 samples WriteHuman passed it where Grubby failed, and the reverse happened on 2. On Originality.ai, ZeroGPT and Winston AI the two were close to even.

Per-sample agreement

Per-sample agreement, September 2026. Counts are out of the 49 shared samples, computed at the 50% human-score threshold. Margin counts the samples only one tool passed, higher count first.
DetectorPassed more oftenMarginBoth passedNeither passed
Pangram 4.0WriteHuman17 to 21515
GPTZero 2026-09-13-baseWriteHuman3 to 0460
Originality.ai turboTied1 to 1470
ZeroGPT defaultTied0 to 0490
Winston AI 5.0Tied0 to 0490
All 5 detectorsWriteHuman17 to 21515

Every count is computed from scores.csv at build time, not transcribed. “Neither passed” counts the texts that defeated both tools; on Pangram that was 15 of 49 here.

Quality and output discipline

Quality and output discipline, September 2026.
MeasureWriteHumanGrubby
22%72%
49%46%
98%98%
65%94%
1.021.11
00
00
API defaultAcademic
Websitewritehuman.ai grubby.ai

On per-sample quality scores, WriteHuman scored higher on 7 samples, Grubby on 27, and the two tied on 15. WriteHuman’s output was shorter on 37 of 49 samples; mean length ratios were 1.02 and 1.11, or 2% longer than the input and 11% longer than the input respectively.

Where each one outperforms the other

WriteHuman

  • Pangram: passed 32 of 49 against 17, including 17 samples where Grubby failed, against 2 the other way.
  • GPTZero: passed 49 of 49 against 46, including 3 samples where Grubby failed, against 0 the other way.
  • Winston AI: tied at 49 of 49 but on a higher mean human score, 99.8% to 99.3%. It passed 0 samples where Grubby failed.
  • Naturalness: 49% against 46%.
  • Length: shorter output on 37 of 49 samples, which matters under a word limit.

Grubby

  • Pangram: passed 2 samples where WriteHuman failed, even though it was detected more often overall, passing 17 to 32.
  • Originality.ai: tied at 48 of 49 but on a higher mean human score, 98.2% to 96.1%. It passed 1 sample where WriteHuman failed.
  • ZeroGPT: tied at 49 of 49 but on a higher mean human score, 96.0% to 94.8%. It passed 0 samples where WriteHuman failed.
  • Factual consistency: 72% against 22%.
  • Stance preservation: 94% against 65%.

WriteHuman vs Grubby FAQ

Is WriteHuman better than Grubby?

In the September 2026 HumanizerEval review, WriteHuman cleared all 5 detectors on 32 of the 49 shared samples and Grubby on 17. On Originality.ai, ZeroGPT and Winston AI the two were within 6 points of each other. Quality ran 7 samples to 27, with 15 ties.

Which humanizer is best for Pangram, WriteHuman or Grubby?

WriteHuman. In September 2026, WriteHuman passed Pangram on 32 of the 49 shared samples and Grubby on 17. On 17 samples WriteHuman passed where Grubby failed; the reverse happened on 2. Both failed on 15.

Which humanizer changes the text less?

Grubby. It scored higher on per-sample quality on 27 of 49 samples, against 7. WriteHuman’s outputs averaged 1.02× the input length and passed factual consistency on 22%; Grubby’s averaged 1.11× and passed on 72%. On stance preservation the split was 65% to 94%. WriteHuman produced the shorter output on 37 of 49 samples.

More comparisons