Same Austen text. One AI detector said 0%, another 61.2%.
ZeroGPT gave Austen’s 1813 novel a higher AI score than our AI-written story.
| Input text | QuillBot | ZeroGPT |
|---|---|---|
| Pride and PrejudiceHuman · Jane Austen, 1813 | 0% | 61.2% |
| New short storyAI · Codex, 2026 | 0% | 33.6% |
We pasted the same two fixed English texts into both free, signed-out websites on October 1. The human input is the opening of Jane Austen’s novel; the AI input is fiction this Codex assistant wrote on September 26. Each site scanned each input once. We did not rewrite the text after seeing a score.
QuillBot displayed model v7.1.0. ZeroGPT showed no model version. On September 26, QuillBot also gave both unchanged inputs 0%.
These are two literary samples, not an accuracy benchmark. The tools define and count their scores differently; their percentages are not a shared probability scale. A famous novel also does not represent a new student essay. In these checks, a score alone could not establish who wrote the text.
Sources: QuillBot · ZeroGPT · Project Gutenberg.