Skip to content
November 2026 results: WriteHuman first, Phrasly climbs to fifth. Read the analysis →
HumanizerBench

Head to head · November 2026 cycle

Rephrasy vs StealthGPT

StealthGPT finished 17.0 points ahead of Rephrasy in the November 2026 cycle, 69.47 to 52.50, ranking #2 against #16 in a field of 16. StealthGPT scored higher on 1 of 6 detectors and 6 of 6 writing categories. Rephrasy came out ahead on GPTZero (96.2 vs 71.4), Winston AI (95.3 vs 81.1) and ZeroGPT (75.8 vs 63.6). The bypass-rate gap, 73.9 to 66.0, sits inside both tools' 95% confidence intervals, so treat bypass as a draw.

52.50 /100
overall score
#16 /16
rank -
+16.97
StealthGPT leads →
Detectors
5 – 1
Categories
0 – 6
Components
2 – 3
69.47 /100
overall score
#2 /16
rank -

30 prompts · Last tested · Methodology v1.3.0

Score components

What the overall score is made of: bypass rate weighs 42%, meaning preservation 32%, readability 16%, consistency across categories 10%. Output-quality penalties come off the total. How scoring works →

Bypass rate CIs overlap Rephrasy by 7.9
73.9 66.0
Meaning preservation StealthGPT by 22.4
44.9 67.3
Readability StealthGPT by 25.0
52.7 77.7
Consistency Rephrasy by 8.8
96.8 88.0
Penalties fewer is better StealthGPT by 10.0
−11.0 −1.0

CIs overlap marks a gap that sits inside both tools' published 95% confidence intervals for that component; it may not survive another cycle.

Detector by detector

Share of each tool's outputs that a detector classified as human-written. Rephrasy took 5 of 6.

GPTZero Rephrasy by 24.8
96.2 71.4
Originality.ai Rephrasy by 2.2
31.0 28.8
Copyleaks Rephrasy by 6.7
96.7 90.0
Winston AI Rephrasy by 14.1
95.3 81.1
ZeroGPT Rephrasy by 12.2
75.8 63.6
Pangram StealthGPT by 12.5
48.4 60.9

Writing categories

Category score per writing context: bypass credit only for real rewrites, blended with the category's own meaning preservation and readability. Small categories swing; the prompt count is shown on each row. How category scores work →

Academic Essay Application Essay Blog Post Marketing Copy Discussion Board News Article Academic Essay: Rephrasy 50.9 · StealthGPT 69.7 Application Essay: Rephrasy 48.1 · StealthGPT 63.4 Blog Post: Rephrasy 40.6 · StealthGPT 75.4 Marketing Copy: Rephrasy 47.8 · StealthGPT 60.6 Discussion Board: Rephrasy 66.1 · StealthGPT 67.5 News Article: Rephrasy 46.1 · StealthGPT 65.6
Rephrasy StealthGPT Outer ring = 100
Academic Essay 6 prompts StealthGPT by 18.8
50.9 69.7
Application Essay 6 prompts StealthGPT by 15.3
48.1 63.4
Blog Post 6 prompts StealthGPT by 34.8
40.6 75.4
Marketing Copy 6 prompts StealthGPT by 12.8
47.8 60.6
Discussion Board 3 prompts StealthGPT by 1.4
66.1 67.5
News Article 3 prompts StealthGPT by 19.6
46.1 65.6

Where each one wins

Every measure above that one tool won outright, largest margin first within each group.

Rephrasy

7 wins
  • Consistency Score 96.8 vs 88.0
  • Bypass rate Score 73.9 vs 66.0
  • GPTZero Detector 96.2 vs 71.4
  • Winston AI Detector 95.3 vs 81.1
  • ZeroGPT Detector 75.8 vs 63.6
  • Copyleaks Detector 96.7 vs 90.0
  • Originality.ai Detector 31.0 vs 28.8

StealthGPT

10 wins
  • Penalties Score −1.0 vs −11.0
  • Readability Score 77.7 vs 52.7
  • Meaning preservation Score 67.3 vs 44.9
  • Pangram Detector 60.9 vs 48.4
  • Blog Post Category 75.4 vs 40.6
  • News Article Category 65.6 vs 46.1
  • Academic Essay Category 69.7 vs 50.9
  • Application Essay Category 63.4 vs 48.1
  • Marketing Copy Category 60.6 vs 47.8
  • Discussion Board Category 67.5 vs 66.1

Same prompt, both outputs

Both tools were given this AI-written passage in the November 2026 cycle. Below it is what each one returned. This prompt was picked by a fixed rule, not by how the tools scored.

Personal Statement · Application Essay · written by gemini-3-8-flash · 328 words

Standard economic theory told me that humans are rational optimizers—cold, calculating machines operating within a frictionless world of supply and demand. However, watching the world around me, particularly during the economic turbulence of recent years, revealed a starkly different reality. People panic-buy during crises, succumb to the sunk cost fallacy in failing investments, and systematically choose short-term gratification over long-term financial security. It was this glaring disconnect between the elegance of classical models and the messy reality of human decision-making that first ignited my passion for behavioral economics.

During my undergraduate studies in economics, I found myself constantly questioning the *Homo economicus* assumption. While I respected the predictive power of quantitative modeling, the numbers felt hollow without the human element. My intellectual turning point came when I encountered Richard Thaler’s concept of "mental accounting." Suddenly, abstract financial behaviors—such as treating tax refunds differently from regular paychecks—made intuitive sense. I realized that to truly fix economic inefficiencies, we cannot merely assume how people *should* behave; we must study how they *actually* behave.

This fascination led me to research the intersection of choice architecture and public policy. I became deeply interested in how subtle interventions, or "nudges," could yield profound societal benefits—from increasing retirement savings rates to encouraging sustainable energy consumption. Behavioral economics represents the democratization of economic theory, humanizing data through the lens of psychology, neuroscience, and sociology.

At your institution, I want to bridge the gap between theoretical research and applied policy design. I am particularly drawn to your focus on experimental economics and empirical field research, where I hope to investigate how cognitive biases impact wealth inequality and financial inclusion.

I do not want to study economics merely to interpret markets from a detached perspective; I want to understand the psychological architecture that drives them. A graduate degree in behavioral economics will equip me with the interdisciplinary tools required to design policies, products, and systems that work with the grain of human nature, rather than against it.

Show the full input

323 words (input 328)

For years, I thought that I wanted to go into economics, because economics attempts to answer questions about human decisions. Then I realized that it wasn't really about human decisions, it was all about the phenomenon of supply and demand. Supply and demand are really helpful to understand how markets work, but they can't really answer questions about why humans decide anything at all. Yet, the study of economics is so pervasive and successful, that I wondered if there could be a way to incorporate "how we make decisions" into a system.

In my undergraduate studies, I delved into the burgeoning literature around decision theory, the psychology of how people view risk and reward, and the simple mathematical models that ascribe numbers to the former and the latter. Then I encountered actual people, who proved to be stubborn and messy, consistently making decisions that did not adhere to the models. Rather than assuming people are rational and that they only need more and better information, I found the best way to understand people's decisions is to put them under the microscope.

That curiosity brought me into the world of behavioral economics, where economics and psychology meet in a quest to understand and explain human behaviors that, at first glance, seem strange and irrational. Such subtle nudges affect how we save for the future, spend our money, and make choices in our day-to-day life. Behavioral economics brings about a new theoretical model of how people behave.

Behavioral economics appeals to me because it steps beyond the imaginary blue box of traditional economics. I hope to study human behavior because each individual is more than a statistic, existing in a complex web of cultural and social realities that traditional economics neglects.

A graduate degree, I believe, will train me to measure and predict behaviors and then translate my research into solutions. That lets me respond to fundamental questions: In which contexts do these behaviors arise?

Show the full output
  • GPTZero1.00passed
  • Pangram0.39caught
  • Winston AI1.00passed
  • ZeroGPT0.84passed
  • Copyleaks1.00passed
  • Originality.ai0.10caught

Passed 4 of 6 detectors.

Every Rephrasy run in November 2026 →

343 words (input 328)

Throughout history, people have behaved in ways that defy conventional economic thinking. They make emotional and illogical decisions about their finances, cling to investments that are failing (fall victim to the sunk-cost fallacy), and choose immediate gratification over financial stability.

Conventional economic theory paints a picture of a world without friction, in which supply and demand are governed by logical optimizers. But I’ve seen first-hand how far reality strays from theory’s predictions, and this discrepancy is what first attracted me to the field of behavioral economics.

My undergraduate studies in economics provided me with a solid grounding in quantitative modeling, but I was left wanting a deeper understanding of the motivations and decision-making processes of the humans represented by the numbers. Then I came across Richard Thaler’s notion of “mental accounting,” and suddenly disparate concepts came together, and it made sense why people spend tax refunds differently than their paychecks, for example. Instead of simply accepting the Homo economicus assumption, I wanted to delve deeper to better understand the psychology of individual behavior, because in order to devise more efficient economies, we need to understand how people actually behave, not just how we think they should behave.

I am particularly interested in choice architecture and its potential applications in the public policy space. There are a host of applications for “nudges” that can have a dramatic effect on society, increasing retirement savings and encouraging more sustainable energy consumption, for example. Behavioral economics is a field that can truly democratize economics by contextualizing it with insights from other disciplines, including sociology, psychology, and neuroscience.

Rather than studying economics with an arm’s-length understanding of the market, I hope to investigate the psychological forces that give rise to it. With a graduate degree in behavioral economics, I could learn to apply an interdisciplinary understanding of human behavior to the design of policies, systems, and products. Your emphasis on experimental economics and empirical field research would provide me with an opportunity to investigate the impact of cognitive biases on topics such as financial inclusion and wealth inequality.

Show the full output
  • GPTZero0.00caught
  • Pangram0.00caught
  • Winston AI0.93passed
  • ZeroGPT0.77passed
  • Copyleaks1.00passed
  • Originality.ai0.05caught

Passed 3 of 6 detectors.

Every StealthGPT run in November 2026 →

Each chip is one detector's score for that output, from 0 (AI) to 1 (human). A score of 0.50 or higher counts as a pass and shows green; below that the detector caught it and the chip shows red.

Questions people ask

Is Rephrasy better than StealthGPT?

StealthGPT finished 17.0 points ahead of Rephrasy in the November 2026 cycle, 69.47 to 52.50, ranking #2 against #16 in a field of 16. StealthGPT scored higher on 1 of 6 detectors and 6 of 6 writing categories. Both tools ran the same 30 prompts, scored by the same 6 detectors, under methodology v1.3.0; every input, output and verdict is published.

Which bypasses AI detectors better, Rephrasy or StealthGPT?

Rephrasy posted the higher bypass rate, 73.9 vs 66.0, and scored higher on 5 of 6 detectors. Rephrasy led on GPTZero, Originality.ai, Copyleaks, Winston AI and ZeroGPT; StealthGPT led on Pangram. The overall bypass gap is inside both 95% confidence intervals, so it is not a reliable difference.

Do Rephrasy and StealthGPT keep the original meaning?

StealthGPT preserved meaning better, 67.3 vs 44.9 on our 0–100 similarity scale. On readability, StealthGPT rated higher, 77.7 vs 52.7. Penalties this cycle: Rephrasy −11.0, StealthGPT −1.0.