Originality.ai AI detector
Vendor Originality AI Inc. Official site originality.ai How this card is sourced
Updated · AI Detector Index editors
AI Detector By Originality.ai, launched in 2022, detects AI text and AI-written code and states coverage of 30+ languages according to the vendor's home page read on 8 November 2026 (originality.ai). Originality.ai shares a code modality with Copyleaks. How this card is sourced explains the rules; the full detector index holds the other twenty-one cards.
What Originality.ai detects
| Field | Value | Source |
|---|---|---|
| Content types detected | Text, code | originality.ai/blog/introducing-originali... |
| Generators covered | 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 | originality.ai |
| Languages | 30 languages stated | originality.ai |
| Input limits | 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. | originality.ai/pricing |
| Free tier | Free, no subscription required, $0: 10K words (100 credits one time) | originality.ai/pricing |
| API | Yes, docs.originality.ai/api-v2-0-new | docs.originality.ai/api-v2-0-new |
| Integrations | Chrome Extension (Google Docs, Writing Replay), Firefox Extension, Google Classroom (via Chrome extension), Moodle Plugin 2.0, API, MCP, Bulk Scan | originality.ai/extensions |
| Data retention | Privacy Policy (Date of Last Amendment: June 17, 2026): When you use our service to check whether text was AI-generated or human-written, the text you input, and the result generated may constitute scan history. Unless you opt out, we may use your scan history and scan results to help train and improve our models. We do not use your personal data or any Student Data to train AI or large language models; the only data used is the scans/input text, and ALL users have the option to opt out. We retain Personal Information only for as long as reasonably necessary. Certain features use OpenAI's API; OpenAI may retain submitted text in abuse-monitoring logs for up to 30 days. | originality.ai/privacy-policy |
| Launch year | 2022 | originality.ai/blog/ai-accuracy |
| Ownership | Originality AI Inc., 64 Hurontario St, Collingwood, Ontario L9Y 2L6, Canada (founder Jon Gillham; co-founder and CTO Conor Watt). Parent company not published. | originality.ai/privacy-policy |
| Independent studies | 7 records | Studies index |
Originality.ai covers two modalities, text and code. The home page names GPT-6 Astra, GPT-6.1 Sol, Claude Fable 5, Claude Opus 5.5, Gemini 3, Gemini 4 Argon, Kimi K3, DeepSeek V4, Grok 4.7 and Llama 4 Maverick and Scout, "and more" (originality.ai, read 8 November 2026). The pricing and Moodle pages carry older model lists; this card uses the home page.
The vendor states 30+ languages from an internal test across 30 languages and publishes no language list (originality.ai, read 8 November 2026). The index records the count as 30.
On edited AI text, the 5.0 launch post reports humanizer detection of 89.52% for 5.0 and 92.38% for 5.0 Max on 420 humanized AI outputs per detector, a set with no human controls, so those figures measure AI detection only (5.0 launch post, read 8 November 2026).
Originality AI Inc. of Collingwood, Ontario, owns the product. Integrations are Chrome and Firefox extensions, Google Classroom through the Chrome extension, a Moodle Plugin 2.0, an API documented at docs.originality.ai, MCP and Bulk Scan (extensions page, read 8 November 2026).
Limits and plans
| Plan | Price | Limits | Checked on |
|---|---|---|---|
| Pro (yearly) | $12.95/month | 200K words (2K credits) per month, 30 Day Scan History, Deep Scan, Bulk Scan, Share Report, Download PDF, upload files (docx, pdf, doc), team management, additional team seats $8.62/month, standard support. | |
| Pro (monthly) | $14.95/month | Same as Pro (yearly); additional team seats $9.95/month. | |
| Enterprise (yearly) | $136.58/month | 1.5M words (15K credits) per month, 1 year scan history, API access, additional team seats $19.04/month, priority support. | |
| Enterprise (monthly) | $179/month | Same as Enterprise (yearly); additional team seats $24.95/month. | |
| Credit costs | 5.0: 1 credit/100 words ($0.01/100 words, $0.10/1,000 words); 5.0 Deep Scan 5 credits ($0.05, $0.50); 5.0 Max 10 credits ($0.10, $1.00); 5.0 Max Deepscan 50 credits ($0.50, $5.00) | 2 credits are used for 100 word checks of AI and Plagiarism. Pay-as-you-go credits expire 2 years after purchase; subscription credits expire at the end of each month. |
Scans need a 100-word minimum, and 1 credit covers 100 words. Input can be pasted, uploaded as docx, pdf or doc, or given as a URL (originality.ai/pricing, read 8 November 2026).
The free tier is 10K words, or 100 credits, granted one time with no subscription (originality.ai/pricing, read 8 November 2026). A compare table on the home page shows a Basic row of 60 credits renewing daily; this card uses the pricing card figure.
Prices below are from the Originality.ai pricing page, read 8 November 2026, in USD.
On retention, the privacy policy amended June 17, 2026 says: "Unless you opt out, we may use your scan history and scan results to help train and improve our models", and that ALL users have the option to opt out (privacy policy, read 8 November 2026).
What the vendor claims
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.
Originality AI Inc. says: "Originality.ai 5.0 and 5.0 Max have 99%+ accuracy" in the accuracy blog of October 25, 2026 (accuracy blog, read 8 November 2026). The 5.0 launch post of the same date gives AI recall of 97.69% (5.0) and 98.00% (5.0 Max) on UChicago Booth AI, 99.33% on Epoch AI, and 90.1% of mixed documents identified by 5.0 Max (originality.ai/blog/introducing-originality-5-0-and-max, read 8 November 2026). The accuracy blog gives 99%+; the launch post gives 97.69% recall on the Booth set; this card quotes both.
On false positives, Originality AI Inc. says: human specificity on PERSUADE 2.0 of 99.954% (5.0) and 99.958% (5.0 Max); on Liang TOEFL 98.901% and 100%; on PELIC 99.940%; and for code, "no false positives were observed among the 150 human-written samples" (originality.ai/blog/introducing-originality-5-0-and-max, read 8 November 2026).
Originality AI Inc. also says: "Use an AI score as one signal, not a final decision" (originality.ai/blog/ai-accuracy, read 8 November 2026). A specificity figure above 99.9% describes a test set, and the writer in front of you was not in that set.
What independent studies found
- AIDI-S-0002Liang 20232023
Originality.AI was one of seven detectors tested (accessed March 15, 2023). Across the seven, the paper reports an average false positive rate of 61.22% on 91 human-written TOEFL essays and near-perfect accuracy on 88 US 8th-grade essays. Per-detector rates appear only in a bar chart (Fig. 1a) and are not stated as numbers.
- AIDI-S-0003Walters 20232023
Originality.ai (detection model 1.4): 98% of 126 documents correct, 0 incorrect, 2% uncertain; 100% correct on both the 42 GPT-3.5 and 42 GPT-4 papers; 95% of 42 human papers correct, 0 incorrect, two human papers uncertain. Tested June 25 to July 12, 2023.
- AIDI-S-0006RAID 20242024
Originality: false positive rate 0.47% / 0.25% / 0.17% / 0.07% at naive thresholds 0.25 / 0.5 / 0.75 / 0.95 on human-written documents (Table 4); accuracy at a fixed 5% FPR 98.6% on open-source chat models (greedy) and 99.9% on open-source non-chat models (greedy) (Table 5); 85.0% with no attack, 96.7% under paraphrase, 9.3% under the homoglyph attack (drop 75.7%) (Table 6).
- AIDI-S-0012Orenstrakh 20232023
Scored 93.86% on 114 human and 70.00% on 20 ChatGPT English coursework submissions, with 7 false positives in 114; accuracy on 10 QuillBot-paraphrased ChatGPT texts fell from 100% to 40% (Tables 6, 8 and 9).
- AIDI-S-0014Jabarian and Imas 20252025
False positive rate 0.0011 at a 0.5 threshold on 1,992 human passages and at most 0.0218 per genre at Youden-optimised thresholds; false negative rate up to 0.4242 on Claude Opus 4 text at the 0.9 threshold and 0.0227 to 0.2073 after StealthGPT rewrites (Tables 3, 6, 4 and B.4).
- AIDI-S-0018Pegoraro 20232023
True positive rate 7.6% on ChatGPT responses and true negative rate 95.0% on human responses in a 20-method sweep on a ChatGPT and human Q&A benchmark; the authors note the tool is prone to classify ChatGPT content as real (Table I and Section II-E).
- AIDI-S-0026Epoch AI 20262026
Originality.ai Turbo 3.0.2 flagged 19 of 495 pre-2022 human passages (3.84%, 16 of them fiction), missed 1 of 297 AI passages from a plain prompt (0.34%) and 53 of 297 (17.85%) written to imitate a named author's style, at its default verdict (Epoch AI results table, June to July 2026).
- AIDI-S-0037Originality.ai vendor compilation2026, vendor benchmark
Vendor compilation (Originality.ai's own, accuracy page and meta-analysis post dated 7 November 2026): 5.0 and 5.0 Max stated at 99%+ (accuracy) with no test-set size; release history gives false positive rates of 0.5% (Lite 1.0.2) and 1.5% (Turbo 3.0.2), both retired September 2026; a vendor-selected list of 17 third-party studies restates one headline figure per study of 83% to 100% for Originality.ai, including 85% base accuracy on RAID (model 2.0 Standard) and 97% overall on Walters 2023 where the paper prints 98%. Compiled by the vendor; omits five index records that scored the product.
Liang 2023 ran seven detectors, Originality.AI among them, on 91 human-written TOEFL essays in March 2023; the average false positive rate across the seven was 61.22%, with per-detector rates only in a bar chart (arXiv 2304.02819, read 8 November 2026).
Walters 2023 found Originality.ai (detection model 1.4) 98% correct on 126 documents between June 25 and July 12, 2023: 100% on both the 42 GPT-3.5 and 42 GPT-4 papers, and 95% of 42 human papers correct with 0 incorrect and two uncertain (De Gruyter PDF, read 8 November 2026).
The RAID benchmark (2024) reports an Originality false positive rate of 0.47%, 0.25%, 0.17% and 0.07% at thresholds 0.25, 0.5, 0.75 and 0.95; accuracy at a fixed 5% FPR of 98.6% on open-source chat models and 99.9% on non-chat models; and 96.7% under paraphrase but 9.3% under the homoglyph attack (arXiv 2405.07940, read 8 November 2026).
Two records were added on 9 November 2026. Orenstrakh 2023 scored Originality.ai at 93.86% on 114 human and 70.00% on 20 ChatGPT English coursework submissions, with 7 false positives in 114 and a fall from 100% to 40% on 10 QuillBot-paraphrased texts (arXiv 2307.07411, read 9 November 2026). Jabarian and Imas 2025 reports a false positive rate of 0.0011 at threshold 0.5 on 1,992 human passages and at most 0.0218 per genre at Youden-optimised thresholds, false negative rates up to 0.4242 on Claude Opus 4 text at the 0.9 threshold and 0.0227 to 0.2073 after StealthGPT rewrites (NBER w34223, read 9 November 2026).
Pegoraro 2023, added 9 November 2026 as the sixth record, ran 20 detection methods on a ChatGPT and human question-and-answer benchmark in March 2023 and records Originality.ai at a 7.6% true positive rate on ChatGPT responses and a 95.0% true negative rate on human responses, with the authors noting the tool "is prone to classify ChatGPT-generated content as real" (arXiv 2304.01487, Table I and Section II-E, read 9 November 2026); no ranking field takes a value from the printed rates.
Epoch AI 2026 joined the list on 9 November 2026 and took the count to 7: Originality.ai Turbo 3.0.2 flagged 19 of 495 pre-2022 human passages (3.84%, 16 of them fiction), missed 1 of 297 plain-prompt AI passages (0.34%) and 53 of 297 passages written to imitate a named author (17.85%), none of which replaces the Walters 2023 zeros in the false positive and false negative fields because neither is lower.
Where Originality.ai appears in the index
Rankings
- AIDI-R-01AI detectors ranked by independent testing Two records added in the fifth wave on 9 November 2026, Pegoraro 2023 and Fleckenstein 2024, changed nine counts and none: Pegoraro 2023 tested ZeroGPT, GPTZero, Copyleaks, Originality.ai, Writer, Content at Scale (BrandWell), DetectGPT, the GPT-2 Output Detector and GLTR, lifting GPTZero to 12, the GPT-2 Output Detector to 10, Copyleaks to 7 and Originality.ai and ZeroGPT to 6; Fleckenstein 2024 tested 289 teachers and no software, so it counts for no card ([Fleckenstein 2024 record](/studies/fleckenstein-2024-teachers/)). Illuminarty, one of four cards added in the same wave, joined with its one image record from Ha and Passananti 2024, so 18 of 72 cards now qualify.
- AIDI-R-02AI detector false positive rates, ranked
- AIDI-R-04AI detectors ranked by languages supported
- AIDI-R-07AI image, voice, video and code detectors ranked by independent studies
- AIDI-R-08AI detectors ranked by published robustness to paraphrased and edited text
- AIDI-R-09AI detector false negative rates, ranked
- AIDI-R-10AI detector accuracy claims vs published results, ranked by the gap
- AIDI-R-11AI detectors ranked by named integration surfaces
- AIDI-R-12AI detectors ranked by the upload formats the vendor states
- AIDI-R-13AI detectors ranked by minimum text length per check
- AIDI-R-14AI detector APIs ranked by published price per 1,000 words
- AIDI-R-15AI detectors ranked by the LMS platforms the vendor names
- AIDI-R-17AI detectors ranked by the generator models the vendor names
- AIDI-R-18AI detectors ranked by study records on non-English and non-native text
- AIDI-R-19AI code detectors ranked by the programming languages the vendor names
- AIDI-R-20AI detectors ranked by launch year, with every change of hands and withdrawal the cards record
Fits these use cases, questions and comparisons
- AIDI-U-01AI detectors for business and managers
- AIDI-U-02AI detectors for school districts
- AIDI-U-03AI detectors for copywriters and agencies
- AIDI-U-04AI detectors for job applications and recruiters
- AIDI-U-05AI code detectors in the index
- AIDI-U-08AI detectors for researchers checking their own manuscripts
- AIDI-U-10AI detectors for grant and proposal reviewers
- AIDI-U-12AI detectors for localisation and translation teams
- AIDI-Q-01Is there a 100% accurate AI detector? What published tests show
- AIDI-C-04GPTZero vs Pangram vs Originality.ai: the three tools one 2025 study ran on the same 9,960 passages
- AIDI-C-08Pangram vs Originality.ai: the 14 fields and the study results side by side
- AIDI-C-09Winston AI vs Originality.ai: the 14 fields and the study results side by side
Originality.ai qualifies for three of the five launch rankings and three of the five added on 9 November 2026: independently tested on 5 studies, false positive rate on the Walters 2023 figure of 0 in 42, and languages on a count of 30 (detectors.json, read 8 November 2026). The free allowance is one time, not monthly, so the card is absent from the free tiers ranking.
Since 9 November 2026 Originality.ai also ranks second in false negative rate on 0 of 84 AI papers missed in Walters 2023, second in robustness to paraphrase on 96.7% under RAID's paraphrase attack, and fourth in the claim vs evidence gap on "99%+" claimed against 100% on 84 papers in Walters 2023.
Originality.ai fits business through the API and team seats, school districts through the Moodle plugin and Google Classroom, agencies through the API, Bulk Scan and site scans, recruiters through the one-time free 10K words on text, and code through the code modality (originality.ai/extensions, read 8 November 2026). The nearest card by study overlap is GPTZero. The home page carries the detector register.
Questions about Originality.ai AI detector
How much does Originality.ai cost?
$12.95/month for Pro billed yearly or $14.95/month billed monthly, each with 200K words per month, per the pricing page read 8 November 2026. Enterprise is $136.58/month yearly or $179/month monthly with 1.5M words.
Is Originality.ai free?
Only once. The pricing page read 8 November 2026 grants 10K words (100 credits) one time with no subscription, and no recurring monthly free allowance exists.
What false positive rate do studies give Originality.ai?
0 of 42 human papers in Walters 2023, with two returned as uncertain, and 0.25% at threshold 0.5 in RAID 2024, read 8 November 2026. The vendor's own figure for 5.0 is 99.954% specificity on PERSUADE 2.0.
Sources and revisions
Sources
- originality.aiRead 8 November 2026
- originality.ai/pricingRead 8 November 2026
- originality.ai/extensionsRead 8 November 2026
- originality.ai/blog/introducing-originali...Read 8 November 2026
- originality.ai/blog/ai-accuracyRead 8 November 2026
- originality.ai/privacy-policyRead 8 November 2026
- docs.originality.ai/api-v2-0-newRead 8 November 2026
- arxiv.org/pdf/2304.02819v3Read 8 November 2026
- degruyterbrill.com/document/doi/10.1515/o...Read 8 November 2026
- arxiv.org/pdf/2405.07940Read 8 November 2026
Revision log
- Card created
- Orenstrakh 2023 and jabarian and imas 2025 records added, 5 study records
- Pegoraro 2023 record added to the study list in the fifth wave2026-11-09: epoch ai 2026 record added in the sixth wave, 7 study records
). Illuminarty, one of four cards added in the same wave, joined with its one image record from Ha and Passananti 2024, so 18 of 72 cards now qualify.: register plate](/img/independently-tested-480.webp)
