StealthGPT: A WriteHuman Alternative for Essays and Papers

Choose a humanizer for your next essay or research draft. StealthGPT brings Super, Standard and Lite together with writing, document tools and a built-in AI detector.

StealthGPT Super ranked second overall, one place above WriteHuman Enhanced. Super scored higher on writing quality and led on Pangram and Originality.ai. WriteHuman passed more GPTZero checks.

  • Super, Standard and Lite
  • Essay and paper workflows
  • Published test data

HumanizerEval · November 2026

StealthGPT Super vs WriteHuman on 50 matched inputs

The November benchmark compares 16 humanizers on the same 50 English inputs. Superhumanizer joined on November 18 and the original results were carried forward. Detector rows show passes. Writing quality scores are automated and provisional.

StealthGPT Super

Super, Quality, rephrase enabled, API

WriteHuman

Enhanced, API

HumanizerEval November 2026 scores and strict passes for StealthGPT Super and WriteHuman
MeasureStealthGPT SuperWriteHuman
Overall score70.34/100 (#2)56.21/100 (#3)
Weighted detector pass rate69.00%60.00%
Writing quality73.48/100 (#3)47.38/100 (#15)
Pangram 4.144/50 (88%)33/50 (66%)
GPTZero 5.0b30/50 (60%)40/50 (80%)
Originality.ai 5.019/50 (38%)0/50 (0%)
Winston AI 5.248/50 (96%)50/50 (100%)
ZeroGPT34/50 (68%)46/50 (92%)
Readability68.80/10043.20/100
Meaning preservation67.20/10041.20/100
Factual consistency78.00/10016.00/100

The overall score combines 70% weighted detector passes and 30% writing quality. Detector weights are 40% Pangram, 30% GPTZero, 20% Originality.ai, 5% Winston AI and 5% ZeroGPT. Winston uses a document Human Score of at least 80. The other detectors use qualifying human labels. HumanizerEval is operated by the StealthGPT team.

StealthGPT vs WriteHuman on academic writing

Super met all three stated detector thresholds on two of three passages. WriteHuman did so on none. We tested the same argumentative essay, laboratory discussion and research-paper passage on November 7, 2026. Every first output was saved before scoring, with no rerolls.

Three matched English academic passages per service. Counts meeting each recorded threshold; not a general success-rate estimate.
Recorded criterionStealthGPT SuperWriteHuman
GPTZero: human-class probability above 50%2 of 31 of 3
Originality.ai 5.0: AI-assisted text below 50%3 of 30 of 3
Pangram 4.1: AI + AI-assisted text below 50%3 of 30 of 3
Same output met all three thresholds2 of 30 of 3

Swipe the table to compare results.

Tested settings: StealthGPT Super, quality mode, plain-text API output; WriteHuman: Standard API · Academic tone · English · Enhanced off. Sources were three original AI-assisted English editing fixtures, including fictional teaching data.

GPTZero 2026-11-04-base, observed Originality.ai model-5.0 and Pangram 4.1. These are detector-specific threshold counts, not interchangeable scores or a Turnitin result. The tested WriteHuman configuration was its Standard API model. Its Premium API and paid web plans advertise Enhanced models and different output options; those configurations were not tested here.

Read the complete student study · Download every original and rewrite · Inspect all detector records

Compare the actual essay openings

Exact excerpts from the same library-hours essay. Full outputs and editorial checks are in the download; these excerpts carry no separate quality score.

Original draft

A university library should test later opening hours before making a permanent change. The strongest reason for a trial is not that every student prefers working at night. It is that a single closing time affects students differently.

StealthGPT Super

There may be good reasons why the library opening hours should be extended and students may well wish this was the case. However, blanket statements about what time the library should close are not helpful.

WriteHuman

Testing later hours is critical to permanent change at a university library. Not every student works best at night. It is important to consider how each individual closing time impacts students.

Read the full WriteHuman review

Follow the tested writing example through the humanizer and detector reports in our hands-on review.

  • Is WriteHuman.ai Undetectable?

    August 2024 review: compare WriteHuman and StealthGPT on a blog-writing task, including detector screenshots and an editorial review.

Why students choose StealthGPT over WriteHuman

Choose StealthGPT when you want a dedicated humanizer inside a broader writing workspace. Super, Standard and Lite give you different levels of rewriting; Stealth Writer, document tools and AI detection help you move from an initial draft to a version you can compare with your sources. Our tested model guide shows the actual differences on academic passages.

The measured reason to start with Super is its detector-focused performance. In our November 7 academic test, Super met all three stated thresholds on two of three passages; WriteHuman did so on none. The student comparison study includes every original and unedited output, exact configurations and editorial findings.

Side-by-Side: StealthGPT vs. WriteHuman

November weighted bypass, five detectors

StealthGPT69.00%HumanizerEval, 50 matched drafts per tool
WriteHuman60.00%HumanizerEval, 50 matched drafts per tool

November academic test, all three thresholds

StealthGPT2 of 3 passagesOne first output per source, no rerolls
WriteHuman0 of 3 passagesStandard API model, academic tone and English; Enhanced was off

Humanizer model choices

StealthGPTSuper, Standard and LiteChoose a substantial rewrite or a lighter edit
WriteHumanStandard API and Premium Enhanced options; web plans offer Enhanced modelsConfiguration and plan matter when comparing results

Student writing workspace

StealthGPTHumanizer, Stealth Writer, Chat with PDF and Photo to AnswersDrafting, source-document and editing tools in one workspace
WriteHumanAI humanizer, AI detector, image detector and output variations by planPublished feature descriptions, not a claim that every tool was tested

API and agent workflows

StealthGPTREST, async jobs, bulk submission and MCPDocumented humanization and content-agent workflows
WriteHumanREST humanization and detection; Standard and Premium API plansSee the linked provider documentation

Billing basis

StealthGPTWeb plans and API access are separate; API charges input plus output with model-specific ratesCheck the plan that matches your actual workflow
WriteHumanAPI bills input words; web plans use request and word limitsWeb-subscription and API allowances are different products

50-draft scores: November 2026 HumanizerEval benchmark. The separate November 7 academic results come from our matched student writing study. Feature and billing descriptions reviewed November 7, 2026 against WriteHuman’s web plans and API plans, StealthGPT plans and StealthAPI.

Three reasons to choose StealthGPT for your next paper

A flagship humanizer with inspectable academic results

Super led on every automated writing-quality measure in the November HumanizerEval comparison with WriteHuman. Its combined quality score was 73.48/100 versus 47.38/100 across 50 matched drafts per tool. The separate November 7 student test includes the complete passages so you can assess the rewrites yourself.

StealthGPT’s Advantage:

Read the measured results and complete outputs, then choose Super, Standard or Lite for the amount of editing your draft needs.

Keep the paper and its sources in one workflow

Use Stealth Writer to develop a draft and its citations, Chat with PDF to work with source documents, and the Humanizer to revise repetitive wording. Move through a paper section by section and compare the rewrite with your original argument.

StealthGPT’s Advantage:

Start with the AI Humanizer, then use the research-paper workflow to check citations, quotations, numbers and terminology.

Use the same service in your app or publishing pipeline

When humanization belongs inside a business workflow, StealthAPI supports explicit model selection, asynchronous jobs and bulk requests. MCP tools connect humanization to compatible assistants and coding agents.

StealthGPT’s Advantage:

The API document-preservation test includes six live calls, complete outputs and a runnable client that keeps approved spans fixed while surrounding prose is rewritten.

Choose a plan for your writing workflow

Choose a plan around the paper lengths and revision volume you expect. WriteHuman’s published web plans range from 600 words per request on Basic to 3,000 on Ultra; its API has different limits, including 2,000 words per call on Standard and higher allowances on Premium. Those current options are more useful than treating every WriteHuman plan as the same small text box.

StealthGPT’s advantage is a connected writing workflow and an explicit choice of humanization models, with published results you can inspect. Compare current web plans for student use and StealthAPI for automation. Web-subscription pricing should not be substituted for API billing.

Frequently Asked Questions

Still have more questions?
Learn more in our support page.

Yes on 33 of 50 tested drafts (66%). HumanizerEval’s November 2026 benchmark used Pangram 4.1. StealthGPT Super passed 44 of the same 50 drafts. See the dated detector results.

Choose your next essay rewrite with evidence

Start with Super, compare the result with your original, and keep your writing, document tools and AI checks in one StealthGPT workspace.