turnitin0

Turnitin0 - A trusted institution providing third-party Turnitin checking services to students worldwide.

Research report

Can Turnitin0's Humanization of GPT-5.6-Sol-Generated Essays Bypass Turnitin AI Detection?

A Large-Scale Benchmark Experiment on AI Detection Evasion

Research ID: TT0-2026-0009 Published July 30, 2026

James Peng

James Peng Turnitin0.com | Lead Editor

3 years of writing experiences on Turnitin AI detection and humanizer.

Cite this research

In a benchmark of 174 humanized essays totaling 204,736 words, Turnitin0 achieved an overall word-level evasion rate of 76.44% against Turnitin AI detector, with Education essays reaching 100% and English essays the lowest at 55.41%.
According to Turnitin0 Research, its humanizer successfully evades Turnitin AI detection for 76.44% of words across 174 essays, with Education and Environmental Science majors achieving near-perfect evasion rates.
  • 174 articles tested
  • Checker: Turnitin
  • LLM: gpt-5.6-sol
  • Humanizer: turnitin0
  • 1 linked dataset
  • Detection screenshots available
  • Published July 30, 2026
Peng, J. (2026). Can Turnitin0's Humanization of GPT-5.6-Sol-Generated Essays Bypass Turnitin AI Detection?. Turnitin0 Research. https://www.turnitin0.com/research/reports/can-turnitin0-s-humanization-of-gpt-5-6-sol-generated-essays-bypass-turnitin-ai-detection

Licensed under CC0 1.0 Universal. You may copy, modify, distribute, and cite this research—including text, figures, tables, and datasets—without asking permission.

Overview

This report presents a benchmark experiment evaluating whether the Turnitin0 humanizer can effectively bypass Turnitin's AI detection system. The study uses a dataset of 174 humanized essays (204,736 words) generated by GPT-5.6-Sol and processed through Turnitin0, spanning 30 majors across five domains. The primary metric is word-level accuracy—the percentage of words correctly identified as human-written by Turnitin's detector.

  • Overall word accuracy is 76.44% (156,497 correct words out of 204,736 total words), indicating that Turnitin0 humanization successfully evades detection for the majority of text.
  • Domain performance varies: Education achieves 100% accuracy (6,750/6,750 words), while Humanities shows the lowest at 70.83% (38,231/53,976 words).
  • Major-level results: Education (100%), Environmental Science (98.04%), and History (94.34%) are the most effective; English (55.41%) and Political Science (55.94%) are the least.
  • Academic level: Undergraduate essays (80.63% accuracy) outperform graduate essays (72.03%).
  • Word count: Longer essays (1,200-word bucket) achieve 78.3% accuracy, compared to 68.39% for 400-word essays and 63.8% for 800-word essays.
  • Dataset: 174 humanized essays (204,736 words) generated by GPT-5.6-Sol and processed through Turnitin0, covering 30 majors across Business, Education, Humanities, Psychology, and STEM.
  • Word count distribution: Essays are grouped into 400-word (49 articles), 800-word (12 articles), and 1,200-word (113 articles) buckets.
  • Academic levels: 84 graduate-level and 90 undergraduate-level essays.
  • Evaluation metric: Word-level accuracy—the proportion of words correctly flagged as human-written by Turnitin's AI detection system.
  • No ESL content was included in the dataset.
  • Scoring unit: accuracy is measured in words. First identify correct paragraphs under the expectation below, then sum the words in those paragraphs for the correct-word count; total words are summed over all counted paragraphs.
  • Correct paragraph: a counted paragraph that was not flagged as AI. Correct words are the words in those paragraphs.
  • References excluded: from a References/Bibliography-style heading onward, those paragraphs are left out of word counts.
  • Limited domain coverage: Education and Psychology have only 6 articles each, while STEM has 86, potentially skewing domain-level comparisons.
  • Total articles: 174
  • Total words: 204,736
  • Correct words: 156,497
  • Word accuracy: 76.44%

This overall result shows that Turnitin0 humanization successfully evades Turnitin AI detection for the majority of text.

  • Education: 100% accuracy (6,750/6,750 words) — highest among domains.
  • Psychology: 90.81% accuracy (7,036/7,748 words).
  • STEM: 78.48% accuracy (78,584/100,129 words).
  • Business: 71.67% accuracy (25,896/36,133 words).
  • Humanities: 70.83% accuracy (38,231/53,976 words) — lowest.
  • Top majors: Education (100%), Environmental Science (98.04%), History (94.34%), Physics (93.83%), Psychology (90.81%).
  • Bottom majors: English (55.41%), Political Science (55.94%), Health Sciences (62.28%), Communications (63.39%), Finance (64.37%).
  • Mid-range majors: Biology (88.33%), Computer Science (86.27%), Pre-Med (87.46%), Criminal Justice (80.77%).
  • Variation is wide: from 55.41% to 100% accuracy across 30 majors.
  • Undergraduate: 80.63% accuracy (84,646/104,980 words) — higher evasion rate.
  • Graduate: 72.03% accuracy (71,851/99,756 words) — lower evasion rate.
  • The difference suggests that undergraduate-level text is easier to humanize effectively.
  • 1,200-word essays: 78.3% accuracy (132,985/169,830 words) — best performance.
  • 400-word essays: 68.39% accuracy (18,513/27,070 words).
  • 800-word essays: 63.8% accuracy (4,999/7,836 words) — lowest.
  • Longer essays tend to achieve higher evasion rates, possibly due to more context for humanization.

Detection report screenshots

Turnitin marks suspected AIGC text in cyan.