AI Humanizer BenchmarkDetectors / Winston AI
Best for Winston AI
In the October 2026 cycle, Winston AI scored 363 outputs from the 11 humanizers in the benchmark and caught 125 (34.4%), the 3rd-most of the seven detectors. The test set has no human-written text, so false positives are not measured here.
Highest average Winston AI human-likelihood this cycle: WriteHuman (94.8), GPTinf (89.1) and HIX Bypass (86.6).
An AI detector and plagiarism checker that scores documents for AI content, plagiarism and readability.
How we read its score: Reports a 0-100 human score; divided by 100.
11 humanizers × 33 prompts = 363 outputs · run October 1, 2026 · methodology v1.1.0
Humanizer ranking against Winston AI
Sorted by the average human-likelihood Winston AI gave each tool's 33 outputs: its bypass rate against Winston AI. Six of the 11 tools had 80% or more of their outputs passed by Winston AI. The top three had 31, 30 and 29 of 33 passed. WriteHuman also ranked first in September 2026.
| Rank | Humanizer | Penalties | Last tested | ||
|---|---|---|---|---|---|
| 1 | 94.8 | 75.06 | -4.0 | ||
| 2 | 89.1 | 80.74 | none | ||
| 3 | 86.6 | 83.87 | none | ||
| 4 | 86.5 | 78.33 | -1.0 | ||
| 5 | 85.2 | 68.06 | -9.0 | ||
| 6 |  UndetectedGPTundetectedgpt.ai | 84.4 | 88.12 | none | |
| 7 | 68.8 | 78.23 | -1.0 | ||
| 8 | 60.0 | 80.20 | none | ||
| 9 | 55.3 | 64.52 | -4.0 | ||
| 10 | 3.0 | 33.60 | -12.0 | ||
| 11 | 0.0 | 41.60 | -4.0 |
11 humanizers · 7 detectors · scores out of 100, higher is better. Click a column to sort; hover a penalty for the breakdown.
Pass rate on Winston AI vs the other detectors
In October 2026, every humanizer's pass rate on Winston AI was within 30 points of its rate on the other detectors.
| Humanizer | Passed Winston AI | Passed the other 6 | Difference | Cycles 30+ pts lower |
|---|---|---|---|---|
| 0% 0/33 | 20% 40/198 | −20.2 | 0 of 2 | |
| 61% 20/33 | 74% 147/198 | −13.6 | 0 of 2 | |
| 67% 22/33 | 78% 154/198 | −11.1 | 0 of 2 | |
| 3% 1/33 | 11% 21/198 | −7.6 | 0 of 2 | |
| UndetectedGPT | 85% 28/33 | 91% 181/198 | −6.6 | 0 of 2 |
| 58% 19/33 | 62% 123/198 | −4.5 | 0 of 2 | |
| 94% 31/33 | 97% 193/198 | −3.5 | 0 of 2 | |
| 85% 28/33 | 82% 163/198 | +2.5 | 0 of 2 | |
| 88% 29/33 | 78% 155/198 | +9.6 | 0 of 2 | |
| 91% 30/33 | 81% 161/198 | +9.6 | 0 of 2 | |
| 91% 30/33 | 81% 160/198 | +10.1 | 0 of 2 |
Passed means a human-likelihood of 0.50 or higher. The other six detectors' verdicts on the same outputs are pooled. Difference is the pass rate on Winston AI minus the pooled rate, in percentage points; negative means Winston AI was harder for that tool.
Winston AI compared with the other AI detectors
All seven detectors scored the same 363 outputs in October 2026. Winston AI gave the same verdict as Originality.ai most often (86.2% of outputs) and as GPTZero least often (65.3%).
| Detector | Caught | Catch rate | Change since Sep 2026 | Unique catches | Same verdict as Winston AI |
|---|---|---|---|---|---|
| 203 of 363 | 55.9% | +11.3 | 64 | 65.3% | |
| 145 of 363 | 39.9% | −19.6 | 13 | 79.1% | |
| 125 of 363 | 34.4% | −12.4 | 8 | – | |
| 99 of 363 | 27.3% | −30.3 | 0 | 86.2% | |
| 95 of 363 | 26.2% | +1.9 | 7 | 80.2% | |
| 71 of 363 | 19.6% | −1.7 | 1 | 77.4% | |
| 67 of 363 | 18.5% | +1.4 | 1 | 80.2% |
Change is in percentage points of catch rate. Unique catches are outputs a detector caught while every other detector passed them. Same verdict is the share of outputs where the detector and Winston AI both caught or both passed the output.
Outputs only Winston AI caught
In October 2026, Winston AI caught 8 outputs that all six other detectors passed: 6.8% of the 118 outputs the others let through. Largest score gap first.
| Prompt | Humanizer | Winston AI score | Lowest other score |
|---|---|---|---|
| Listicle blog · GPT-5.6 Terra | UndetectedGPT | 0.00 | 0.88 |
| Product description · Gemini 2.5 Flash | StealthGPT | 0.01 | 0.87 |
| Listicle blog · GPT-5.6 Terra | WriteHuman | 0.09 | 0.85 |
| Argumentative essay · Gemini 2.5 Flash | UndetectedGPT | 0.01 | 0.67 |
| Listicle blog · Claude Haiku 4.5 | CleverHumanizer | 0.24 | 0.75 |
| Listicle blog · Gemini 3.5 Flash | UndetectedGPT | 0.04 | 0.53 |
| Argumentative essay · Claude Haiku 4.5 | SuperHumanizer | 0.38 | 0.70 |
| News article · Claude Haiku 4.5 | SmartHumanizer | 0.40 | 0.64 |
Scores are human-likelihood from 0 to 1. Every other detector scored these outputs 0.50 or higher; the last column is the lowest of those scores. Each prompt opens its evidence page at that humanizer's output.
Winston AI catch rate by writing type
In October 2026, Winston AI caught business email outputs most often (58%) and academic essay outputs least often (23%).
| Writing type | Winston AI | Other 6 detectors |
|---|---|---|
| Academic essay | 23% 15/66 | 22% 88/396 |
| Application essay | 33% 22/66 | 42% 167/396 |
| Blog post | 45% 30/66 | 30% 120/396 |
| Business email | 58% 19/33 | 47% 93/198 |
| Marketing copy | 33% 22/66 | 29% 115/396 |
| Discussion board | 24% 8/33 | 27% 53/198 |
| News article | 27% 9/33 | 22% 44/198 |
The last column pools the other six detectors' verdicts on the same outputs. By prompt, Winston AI caught the most tools on Listicle blog · GPT-5.6 Terra (10 of 11) and the fewest on Discussion post · Claude Sonnet 5 (2 of 11).
Winston AI by month
| Month | Humanizers | Caught | Catch rate | Avg human-likelihood | Rank by catches |
|---|---|---|---|---|---|
| October 2026 | 11 | 125 of 363 | 34.4% | 64.9 | 3 of 7 |
| September 2026 | 11 | 170 of 363 | 46.8% | 53.2 | 3 of 7 |
Rank 1 is the detector that caught the most outputs that month.
Using these results to pick a humanizer
The ranking on this page uses Winston AI only. It does not account for the other six detectors, for how well a rewrite keeps the original meaning, or for readability. A tool can pass Winston AI while changing facts or reading poorly. The Overall column above and the leaderboard cover those.
Results can shift between cycles as detectors and humanizers update, so check which cycle a figure comes from, and each tool's history on its own page.
Before relying on a tool, run a sample of your own text through it, check the output in Winston AI, and compare it with your original for changed facts or dropped citations.
Frequently asked questions about Winston AI
Can Winston AI detect humanized AI text?
Some of it. In October 2026, Winston AI caught 125 of the 363 humanized outputs it scored (34.4%), the 3rd-most of the seven detectors in the benchmark. The test set has no human-written text, so how often Winston AI flags human writing is not measured here.
Can Winston AI be bypassed?
Often. In October 2026, 238 of the 363 humanized outputs (65.6%) passed Winston AI as human. Six of the 11 humanizers had at least 80% of their outputs passed.
What is the best AI humanizer for Winston AI?
By Winston AI alone, WriteHuman: it had the highest average human-likelihood from Winston AI in October 2026 (94.8 out of 100), and Winston AI caught 2 of its 33 outputs. The leaderboard ranks tools on all seven detectors plus meaning and readability.
Which writing types does Winston AI catch most often?
Business email outputs, at 58% in October 2026. Academic essay outputs were caught least often, at 23%.
How do you test Winston AI?
Each month, every humanizer rewrites the same 33 AI-written texts on its default settings, and Winston AI scores every rewrite. Each result is converted to a human-likelihood from 0 to 1, as described above. A score below 0.50 counts as caught.
Raw test data
A JSONL file with one row per test from the October 2026 cycle. Each row has the text Winston AI scored and the score it returned. To check a result, paste a row's text into Winston AI and compare.
Other detectors
Related
How scores are computed
See the full methodologyOverall score formula
How often the rewritten text is classified as human by the detector panel.
How closely the rewrite preserves the original text's meaning.
Whether the output is clear, fluent, natural writing.
Whether performance holds across all seven writing types or varies widely between them.
Penalties (right) subtract from the composite, and the result never goes below zero.
Possible penalties
- Identical to input
The text came back essentially unchanged; no meaningful rewriting occurred.
-2.0-10.0 - Refusal
The tool declined to process the text, typically due to its own content filter.
-1.0-10.0 - Meaning drift
The rewrite departed substantially from the original's meaning.
-1.0-10.0 - Length inflation
The output ran far longer than the input, diluting the AI signal with added words rather than removing it.
-1.0-10.0 - Length deflation
The output ran far shorter than the input; content was dropped rather than rephrased.
-1.0-10.0
Maximum possible total penalty: -50.0 (off 100).