StealthGPT review

StealthGPT has no free tier; its Pro plan costs $1.00 for three days, then $29.99 a month for 50,000 words with a 1,000-word cap per request, read on its pricing page on 9 November 2026. This StealthGPT review covers what each plan buys, what the vendor states about detectors, and where the tool sits in the full humanizer ranking under one published method.

What does StealthGPT itself claim?

StealthGPT's pricing page says more about detectors than most. It carries a logo strip headed with a claim to beat every AI detector, a question in its FAQ about Turnitin and GPTZero answered with a yes, and a banner announcing a new model it says beats Pangram and other major detectors, all read 9 November 2026. This site does not repeat those claims as findings. Turnitin's detector is not available to the public, so no page here states a Turnitin pass or fail.

What Turnitin does publish is relevant. Turnitin states that its AI writing checker helps identify when AI bypassers, which it also calls humanizers, may have been used in a submission, alongside AI-paraphrased text (Turnitin, AI writing detection page, read 9 November 2026). GPTZero states that its model shields against common evasion methods, paraphrasing among them, and keeps its false positive rate at no more than 1 percent when evaluating AI against human text (GPTZero technology page, read 9 November 2026). A humanizer and the detectors it names are moving targets for each other, and any reading is true for its date.

What does the research say about paraphrased text?

Weber-Wulff and colleagues tested 14 AI text detectors and found every one below 80 percent accuracy, with accuracy falling further on paraphrased and machine-translated text (Weber-Wulff et al. 2023, International Journal for Educational Integrity, read 9 November 2026). Krishna and colleagues showed that a paraphrase model dropped DetectGPT's detection rate from 70.3 percent to 4.6 percent at a 1 percent false positive rate (Krishna et al. 2023, arXiv, read 9 November 2026). Paraphrasing moves detector readings; the studies agree on that.

What those studies do not measure is whether the rewrite kept the writer's claims. A rewrite that turns "no effect" into "a small effect" has passed a detector and failed the reader. That is why the method on this site weights meaning kept at 35 points of 100, errors added at 20, detector pass rate at 30 and limits and access at 15. No detector score should be the only evidence in a misconduct case.

What this StealthGPT review draws on

Everything above comes from the vendor's pricing page, Turnitin's and GPTZero's published pages, and the two studies linked. The run figures, when published, sit on the method page and in this review, with the raw outputs free to download.

What does StealthGPT change in your text?

The pricing page names no rewrite modes. Its FAQ mentions a "Samurai engine" without describing it, read 9 November 2026, so there is no documented default and strongest pair to run as two rows. The site's protocol runs each tool on its default setting and, where a stronger documented mode exists, that mode as a second row; for StealthGPT the run uses whatever the interface shows on the run date and records the name in the raw files.

What a humanizer in this category does is paraphrase: substitute words, reorder clauses, vary sentence length and shift tone. The 1,000-word cap shapes how that happens in practice. A 1,500-word essay goes through as two requests, and the two halves are rewritten without sight of each other, so a reader should check the join for a repeated point or a changed term. The reader-side costs the method rates are claims dropped, claims invented, grammar errors introduced per 100 words, and readability change. Those figures appear here once a run is published.