Best AI DetectorsEvery reading on the record
Subject

AI detector test runs

Latest run
Run date
9 November 2026
Samples
180

Every number on this site comes from a dated run. Run RUN-2026-11-A started on 9 November 2026 and is in progress: 120 of its 180 samples are scored on GPTZero and Originality.ai, and every other cell shows a test in progress tag until its readings are in.

Page updated: .

Run log

Newest first. Each row keeps its files, so a past citation still resolves.

One row per run. Versions are as reported by each tool on the run date.
RunDateSamplesDetectors and versionsFiles
RUN-2026-11-A
Latest
120 of 180100 human, 20 AI so far
  • GPTZero 2026-11-04-base
  • Pangram Test in progress
  • Originality.ai model-5.0
  • Copyleaks Test in progress
  • Turnitin Test in progress
  • ZeroGPT Test in progress
  • QuillBot Test in progress
  • Scribbr Test in progress
Readings CSVWhat each column means

Latest run, in summary

Run RUN-2026-11-A: every detector across all 180 samples, then the counts behind each score.

Counts are out of the real bases. Access is plan checks passed.
PlaceToolVersionOverall scoreFalse alarms, of 100Raw AI caught, of 60Edited AI caught, of 20Access
1GPTZero2026-11-04-baseScore pending0 of 10020 of 20Test in progressScored at the end
2PangramTest in progressTest in progressTest in progressTest in progressTest in progressScored at the end
3Originality.aimodel-5.0Score pending0 of 10020 of 20Test in progressScored at the end
4CopyleaksTest in progressTest in progressTest in progressTest in progressTest in progressScored at the end
5TurnitinTest in progressTest in progressTest in progressTest in progressTest in progressScored at the end
6ZeroGPTTest in progressTest in progressTest in progressTest in progressTest in progressScored at the end
7QuillBotTest in progressTest in progressTest in progressTest in progressTest in progressScored at the end
8ScribbrTest in progressTest in progressTest in progressTest in progressTest in progressScored at the end

How to cite a run

  1. Name the run ID and the run date, not just the site.
  2. Quote counts with their base: "00 of 100 human samples", never a bare percentage.
  3. Link the run row on this page, so readers can download the same file.
Citation
Best AI Detectors. AI detector test run RUN-2026-11-A, 9 November 2026. 180 samples. https://bestaidetectors.com/test-runs/#RUN-2026-11-A

In a sentence: In run RUN-2026-11-A, GPTZero and Originality.ai each flagged 0 of 100 human texts as AI (Best AI Detectors, 9 November 2026).