Winston AI vs Originality.ai: the 14 fields and the study results side by side
Originality.ai carries 6 study records in the index and Winston AI 2, and one record scored both on the same benchmark: RAID 2024, where Winston AI's false positive rate was 0.55% and Originality.ai's 0.25% at threshold 0.5 on human-written documents, and both reached between 97% and 99% at a fixed 5% false positive rate on open-source chat models (arXiv 2405.07940, Tables 4 and 5, read 8 November 2026). The two cards are read side by side below on the same 14 fields, with the vendor's wording and the page it came from.
Updated Fields checked 8 November 2026
Short answer
On RAID 2024, the one benchmark that scored both, Originality.ai recorded a 0.25% false positive rate and Winston AI 0.55% at threshold 0.5; at a fixed 5% false positive rate Originality.ai scored 98.6% and Winston AI 97.2% on open-source chat models, and 99.9% against 68.2% on non-chat models with greedy decoding (arXiv 2405.07940, Tables 4 and 5). Winston AI's other record is Weber-Wulff 2023 (36 of 54 correct, 0 false accusations in 18 human documents); Originality.ai adds Walters 2023 (0 of 42 human papers flagged), Liang 2023, Orenstrakh 2023, Jabarian and Imas 2025 and Pegoraro 2023. Winston AI states 11 named languages, image and handwriting checks and plans from $18 a month; Originality.ai states 30 languages without a list, code detection and plans from $14.95 a month (gowinston.ai/pricing, originality.ai/pricing, read 8 November 2026). No detector score should be the only evidence in a misconduct or hiring decision.
How do Winston AI and Originality.ai compare on the 14 fields?
| Field | Winston AI | Originality.ai |
|---|---|---|
| 1. Name, vendor, official domain | Winston AI, Winston AI inc., gowinston.ai | AI Detector By Originality.ai, Originality AI Inc., originality.ai |
| 2. Modalities detected | text, image, handwriting (OCR) | text, code |
| 3. Generators covered | ChatGPT, Claude, Google Gemini, LLaMA, "many others"; images from Midjourney, DALL-E, Stable Diffusion, Nano Banana, ChatGPT Image, Adobe Firefly | GPT-6 Astra (Sol, Luna, Terra), GPT-6.1 Sol, Claude Fable 5, Claude Opus 5.5, Gemini 3, Gemini 4 Argon, Kimi K3, DeepSeek V4 (Flash and Pro), Grok 4.7, Llama 4 Maverick and Scout, "and more" |
| 4. Languages supported | 11: English, French, Spanish, Portuguese, German, Dutch, Polish, Italian, Indonesian, Romanian, Chinese (Simplified) | 30 stated ("30+ languages"); no list published |
| 5. Input limits | Up to 200,000 characters per scan on paid plans (Free row blank); unlimited scans per hour on paid plans; paste or upload .docx, .png and .jpg; file size not published | 100-word minimum; 1 credit = 100 words; paste, upload (docx, pdf, doc) or URL; full site scans; maximum words per scan and file size not published |
| 6. Free tier | Free plan at $0 a month with 2,000 credits for a 14-day trial, not recurring; API registration gives 2,500 credits | Free, no subscription, $0: 10K words (100 credits) one time |
| 7. Paid plans | Essential $18 a month (100,000 credits), Advanced $29 (200,000 credits, 5 team members), Elite $49 (500,000 credits, unlimited members); annual $60, $96 and $156 a year; a WINSTON50 promo code shown at half price; Enterprise by contact | Pro $14.95 a month or $12.95 billed yearly (200K words a month); Enterprise $179 a month or $136.58 billed yearly (1.5M words, API access); credits 1 per 100 words, 5.0 Max 10 credits, Deep Scan 5 credits |
| 8. API | Yes, docs.gowinston.ai | Yes, docs.originality.ai (Enterprise plan) |
| 9. Integrations | Chrome Extension, Google Classroom, WordPress, Zapier, API, GPT apps in ChatGPT | Chrome Extension (Google Docs, Writing Replay), Firefox Extension, Google Classroom via the Chrome extension, Moodle Plugin 2.0, API, MCP, Bulk Scan |
| 10. Vendor accuracy and false positive claims | "Winston ranks first on RAID", "100% accuracy on the Chicago Booth benchmark", "zero false positives across 27,743 human documents"; "99.89% of AI text detected at a 1% false-positive rate"; home page headline "the only AI detector with a 99,87% accuracy rate" | "Originality.ai 5.0 and 5.0 Max have 99%+ accuracy" (blog, 25 October 2026); AI recall 97.69% and 98.00% on UChicago Booth AI, 99.33% on Epoch AI; human specificity 99.954% and 99.958% on PERSUADE 2.0; historic false positive rates 1.5%, 0.5% and 2.4% for older models |
| 11. Independent studies | 2 records: Weber-Wulff 2023, RAID 2024 | 6 records: Liang 2023, Walters 2023, RAID 2024, Orenstrakh 2023, Jabarian and Imas 2025, Pegoraro 2023 |
| 12. Data retention and privacy statement | Privacy policy (17 July 2026): submitted text and images are collected and stored; deleting a report deletes its content from active storage; deleting the account deletes all submitted content; no training on individually identifiable submissions without consent | Privacy policy (17 June 2026): scan history and results may train the models unless the user opts out; no personal or student data used for training; retained only as long as reasonably necessary; some features call OpenAI's API, which may keep text in abuse logs for up to 30 days |
| 13. Launch year and ownership | Not stated; Winston AI inc., no parent company stated | 2022; Originality AI Inc., Collingwood, Ontario, founder Jon Gillham; parent not published |
| 14. Rankings the card appears in | Independently tested, false positive rate, languages, input limits, false negative rate, claim vs evidence gap, platform integrations, file types | Independently tested, false positive rate, languages, adversarial robustness, false negative rate, claim vs evidence gap, studies by content type, platform integrations, file types |
Which studies tested both Winston AI and Originality.ai?
One record. RAID 2024, the shared benchmark by Dugan and colleagues, scored twelve detectors on more than six million generations across eleven generators, eight domains and eleven adversarial attacks, and printed each detector's false positive rate at four naive thresholds and its accuracy at a fixed 5% false positive rate (arXiv 2405.07940, read 8 November 2026). Both vendors cite RAID on their own pages: Winston AI says it "ranks first on RAID" and Originality.ai cites the benchmark in its accuracy blog; the figures below are the paper's own, not the live leaderboard's. Winston AI's other record, Weber-Wulff 2023, did not test Originality.ai; Originality.ai's other five records did not test Winston AI.
| Source | Winston AI | Originality.ai |
|---|---|---|
| RAID 2024, false positive rate on human-written documents at thresholds 0.25, 0.5, 0.75 and 0.95 (Table 4) | 0.75%, 0.55%, 0.38%, 0.21% | 0.47%, 0.25%, 0.17%, 0.07% |
| RAID 2024, accuracy at a fixed 5% FPR, open-source chat models, greedy decoding (Table 5) | 97.2% | 98.6% |
| RAID 2024, accuracy at 5% FPR, open-source non-chat models, greedy decoding (Table 5) | 68.2% | 99.9% |
| RAID 2024, accuracy at 5% FPR, open-source non-chat models with sampling (Table 5) | 29.5% | Not printed in the index record |
| RAID 2024, accuracy under attacks (Table 6) | Not printed in the index record | 85.0% with no attack, 96.7% under paraphrase, 9.3% under the homoglyph attack (drop 75.7%) |
| Weber-Wulff 2023, 54 documents | 36 of 54 correct (67%, rank 5 of 14); 0 false accusations in 18 human documents; 12 of 36 AI documents undetected | Not tested |
| Liang 2023, 91 TOEFL and 88 US essays | Not tested | One of seven detectors averaging a 61.22% false positive rate on TOEFL essays; per-detector rates in a chart only |
| Walters 2023, 126 documents | Not tested | 98% correct (model 1.4), 0 of 42 human papers flagged, 2 uncertain; 100% on both GPT-3.5 and GPT-4 papers |
| Orenstrakh 2023, 164 submissions | Not tested | 93.86% on 114 human submissions, 7 false positives; 100% to 40% on 10 ChatGPT texts after QuillBot |
| Jabarian and Imas 2025, 9,960 passages | Not tested | False positive rate 0.0011 on 1,992 human passages at threshold 0.5; false negative rate up to 0.4242 on Claude Opus 4 at 0.9; 0.0227 to 0.2073 after StealthGPT rewrites |
| Pegoraro 2023, ChatGPT and human Q&A benchmark (Table I) | Not tested | True positive rate 7.6%, true negative rate 95.0% |
Results as each study reports them: arXiv 2405.07940, arXiv 2306.15666, arXiv 2304.02819, De Gruyter 10.1515/opis-2022-0158, arXiv 2307.07411, NBER w34223, arXiv 2304.01487, read 8 and 9 November 2026. "Not tested" means the paper did not run that tool; "not printed in the index record" means the paper's cell is not carried on the card.
What do the two cost over 12 months?
| Plan | Monthly as stated | Annual as stated | Inclusions as stated | Checked on |
|---|---|---|---|---|
| Winston AI Free | $0 /month | 2,000 credits for a 14-day trial, not recurring | 8 November 2026 | |
| Winston AI Essential | $18 /month (shown struck through beside $9 with promo code WINSTON50) | $10.00 /month billed annually, $60.00/year ($5.00 with the code) | 100,000 credits a month; AI, image and deepfake detection, plagiarism, writing feedback, OCR, PDF reports | 8 November 2026 |
| Winston AI Advanced | $29 /month ($14.50 with the code) | $16.00 /month billed annually, $96.00/year ($8.00 with the code) | 200,000 credits a month; up to 5 team members | 8 November 2026 |
| Winston AI Elite | $49 /month ($24.50 with the code) | $26.00 /month billed annually, $156.00/year ($13.00 with the code) | 500,000 credits a month; unlimited team members | 8 November 2026 |
| Originality.ai Free | $0 | 10K words (100 credits) one time | 8 November 2026 | |
| Originality.ai Pro | $14.95/month billed monthly | $12.95/month billed yearly, $155.40 a year | 200K words (2K credits) a month, Bulk Scan, 30-day scan history, team seats $8.62 or $9.95 a month | 8 November 2026 |
| Originality.ai Enterprise | $179/month billed monthly | $136.58/month billed yearly, $1,638.96 a year | 1.5M words (15K credits) a month, API access, 1-year scan history | 8 November 2026 |
| Originality.ai credits | 1 credit per 100 words ($0.01 per 100 words); 5.0 Max 10 credits; Deep Scan 5 credits | Pay-as-you-go credits expire 2 years after purchase | 8 November 2026 |
Prices as shown on gowinston.ai/pricing and originality.ai/pricing, checked 8 November 2026; Originality.ai names USD, Winston AI shows the $ symbol only. Winston AI counts credits and Originality.ai counts words, so the first paid tiers (100,000 credits against 200,000 words) are not like for like.
One of the comparisons in the compare hub. The cards are Winston AI and Originality.ai, and the home board answers which AI detectors exist and who has tested them? Both vendors cite RAID on their own pages, which is why the shared record is read first.
Which card fits which reader?
- Choose Originality.ai ifthe RAID figures decide it: 0.25% false positives at threshold 0.5 against Winston AI's 0.55%, and 99.9% against 68.2% on non-chat models with greedy decoding (arXiv 2405.07940, Tables 4 and 5).
- Choose Originality.ai ifyou want the longer record: six studies, including 0 of 42 human papers flagged in Walters 2023 and 0.0011 of 1,992 in Jabarian and Imas 2025 (De Gruyter, NBER w34223).
- Choose Winston AI ifyou check images or scanned handwriting: image detection and OCR are stated on the pricing page; Originality.ai states text and code (gowinston.ai/pricing, originality.ai, read 8 November 2026).
- Choose Winston AI ifyou want the languages named: 11 are listed, including Polish, Indonesian and Romanian; Originality.ai states 30 without a list (gowinston.ai/ai-content-detector, read 8 November 2026).
- Choose Originality.ai ifyou check code, URLs or whole websites: all three are stated on its card and none on Winston AI's (originality.ai, read 8 November 2026).
- Choose Winston AI ifyour API budget is small: registration gives 2,500 free credits and the documented API comes with every plan, while Originality.ai's API is on the Enterprise plan at $136.58 to $179 a month (gowinston.ai/ai-content-detection-api, originality.ai/pricing, read 8 November 2026).
- Choose Winston AI ifyou do not want submissions used for training: its policy says no training on individually identifiable submissions without consent, while Originality.ai's allows scan history to train models unless the user opts out (gowinston.ai/privacy-policy, originality.ai/privacy-policy, read 8 November 2026).
How does the evidence compare, tool by tool?
Winston AI
- Register ID
- AIDI-D-0005
- Study records
- 2 in the index: Weber-Wulff 2023, RAID 2024
- Shared with Originality.ai
- 1: RAID 2024
- Vendor accuracy claim
- Winston ranks first on RAID, achieved 100% accuracy on the Chicago Booth benchmark, and recorded zero false positives across 27,743 human documents in published evaluations. 99.89% of AI text detected at a 1% false-positive rate, including adversarial attacks. (Headline elsewhere on the home page: The only AI detector with a 99,87% accuracy rate.)
- Free tier
- Free: $0 /month, 2,000 credits / 14 day trial (not recurring). API: register for free and receive 2,500 credits.
- Languages
- 11 stated
Originality.ai
- Register ID
- AIDI-D-0004
- Study records
- 7 in the index: Liang 2023, Walters 2023, RAID 2024, Orenstrakh 2023, Jabarian and Imas 2025, Pegoraro 2023, Epoch AI 2026
- Shared with Winston AI
- 1: RAID 2024
- Vendor accuracy claim
- Originality.ai 5.0 and 5.0 Max have 99%+ accuracy (accuracy blog, October 25, 2026). 5.0 launch post: AI recall UChicago Booth AI 97.69% (5.0) / 98.00% (5.0 Max), Epoch AI 99.33%; 5.0 Max identified 90.1% of mixed documents.
- Free tier
- Free, no subscription required, $0: 10K words (100 credits one time)
- Languages
- 30 stated
How this comparison is built
- Cards compared
- Winston AI and Originality.ai
- Fields
- The 14 fields of every card, in card order
- Shared study records
- 1: RAID 2024
- Fields checked
- 8 November 2026
- Records in the studies index
- 35 records
Both columns are drawn from the same 14 fields in the index data file, read on 8 November 2026 from the vendor page recorded on each card, and from the studies index, where each result is copied as the paper reports the figure, with the base stated. RAID 2024 is the one record that scored both tools on the same benchmark, so its rows sit first in the evidence table; the single-tool records follow with "Not tested" in the other column. Each vendor's own RAID statement stays in the claims row and never in the evidence table. The rules, the recheck schedule (next pass 8 December 2026) and the correction route are on the methodology page; the claim vs evidence gap ranking sets each vendor's headline figure beside its best study result.
Where each figure comes from
Values as each vendor states them on gowinston.ai, gowinston.ai/ai-content-detector, gowinston.ai/pricing, gowinston.ai/privacy-policy, originality.ai, originality.ai/pricing, originality.ai/privacy-policy and the Originality.ai 5.0 post, all read 8 November 2026. An empty cell means the card has no published value for the field.
Questions about Winston AI and Originality.ai
Which study tested both Winston AI and Originality.ai?
RAID 2024, the shared benchmark by Dugan and colleagues: false positive rate 0.55% for Winston AI and 0.25% for Originality.ai at threshold 0.5, and 97.2% against 98.6% accuracy at a fixed 5% false positive rate on open-source chat models (arXiv 2405.07940, Tables 4 and 5, read 8 November 2026).
Which has the lower false positive rate?
On RAID, Originality.ai at every threshold (0.47% to 0.07% against 0.75% to 0.21%). Both have a 0 result elsewhere: Winston AI 0 of 18 in Weber-Wulff 2023, Originality.ai 0 of 42 in Walters 2023 (arXiv 2306.15666, De Gruyter).
Does Winston AI rank first on RAID?
That is the vendor's statement about the live leaderboard (gowinston.ai, read 8 November 2026). The RAID paper of 2024 prints Winston AI at 97.2% and Originality.ai at 98.6% on open-source chat models at 5% FPR; the index records the paper, not the leaderboard.
Which supports more languages?
Originality.ai states 30 and Winston AI names 11; only Winston AI publishes the list (originality.ai, gowinston.ai/ai-content-detector, read 8 November 2026).
Which costs less per month?
Originality.ai Pro is $14.95 a month, or $12.95 billed yearly, for 200,000 words; Winston AI Essential is $18 a month, or $10 billed annually, for 100,000 credits (originality.ai/pricing, gowinston.ai/pricing, checked 8 November 2026). Credits and words are different units.
Do both have an API?
Yes. Winston AI documents one at docs.gowinston.ai with 2,500 credits on registration; Originality.ai's is included in the Enterprise plan (gowinston.ai/ai-content-detection-api, originality.ai/pricing, read 8 November 2026).
Which detects images, and which detects code?
Winston AI states image detection and OCR; Originality.ai states code detection (gowinston.ai/pricing, originality.ai, read 8 November 2026).
Sources and revisions
Sources
- Winston AIgowinston.ai
- Winston AI detector pagegowinston.ai/ai-content-detector
- Winston AI pricinggowinston.ai/pricing
- Winston AI APIgowinston.ai/ai-content-detection-api
- Winston AI privacy policygowinston.ai/privacy-policy
- Originality.aioriginality.ai
- Originality.ai pricingoriginality.ai/pricing
- Originality.ai privacy policyoriginality.ai/privacy-policy
- Originality.ai 5.0 postoriginality.ai/blog/introducing-originali...
- Dugan et al. 2024 (RAID)arxiv.org/pdf/2405.07940
- Weber-Wulff et al. 2023arxiv.org/pdf/2306.15666
- Liang et al. 2023arxiv.org/pdf/2304.02819v3
- Walters 2023degruyterbrill.com/document/doi/10.1515/o...
- Orenstrakh et al. 2023 (read 9 November 2026)arxiv.org/abs/2307.07411
- Jabarian and Imas 2025 (read 9 November 2026)nber.org/papers/w34223
- Pegoraro et al. 2023 (read 9 November 2026)arxiv.org/abs/2304.01487
