How precise is Gradr.ai?

Gradr.ai is built to detect AI-written text with the highest possible certainty and as few false accusations as possible.

97.7%

Efficiency

0.4%

False positives

The best AI detection in English

Efficiency

Gradr.ai Ultimate

97.77 %

Gradr.ai Standard

94.26 %

Pangram

89.3 %

Efficiency is the share of AI-generated texts that are recognized as AI.

False positives

Gradr.ai Ultimate

0.4 %

Gradr.ai Standard

0.5 %

Pangram

0.5 %

False positives are the share of texts written by humans that are wrongly flagged as AI.

How AI detection works

  1. 1

    Check the texts

    Paste in or upload a text yourself. If your school has an integration with its learning platform (LMS), submitted texts are checked automatically.

  2. 2

    Get a list of suspicious texts

    Gradr.ai shows you which texts have signs of AI. You only need to follow up on those, not on every text.

  3. 3

    Follow up

    For schools: Follow our guide for help interpreting the result and how to confirm it.

    For exams or at university: Hold an oral conversation or presentation with the students whose texts are suspicious, and assess them on their competence instead of starting a formal cheating case.

Why you can trust the result

Made especially for English

We have developed our own AI detection models for English, trained on large amounts of English text. That is why Gradr.ai works better in English than generic alternatives.

Catches the latest AI models

Works on text from ChatGPT, Claude, Gemini and others. Also when AI text is mixed with human writing, or rewritten with paraphrasing tools to hide it.

Safety in numbers

A single short text or paragraph flagged as AI-generated on its own can be a mistake. But the chance that several flags are all wrong drops quickly the more AI flags there are.

The risk of a false accusation climbs fast with use

50 texts

22.2 %

100 texts

39.4 %

139 texts

50.1 %

200 texts

63.3 %

300 texts

77.8 %

500 texts

91.8 %

1000 texts

99.3 %

Chance that at least one honest student is wrongly flagged, by number of student texts checked

But more AI flags lower the chance of a wrong assessment

1 flag

1 in 200 (0.5 %)

2 flags

1 in 40,000 (0.0025 %)

3 flags

1 in 8 million (0.0000125 %)

4 flags

1 in 1.6 billion (0.00000... %)

Chance that every AI flag is wrong, by number of flagged texts / paragraphs (relative scale from 0 % to 0.5 %)

Clear advice, not just a number

Every result comes with a recommendation, so you know when you can be confident and when you should double-check.

  • Clear signs of AI

    Large parts of the text are flagged. You can be confident in the result.

  • Likely or possible signs of AI

    Some parts are flagged. Our guide helps you double-check before you conclude.

  • Some uncertain parts

    Only worth a closer look if you have zero tolerance for AI use.

  • No signs of AI

    No follow-up needed.

Contact us

Interested in using Gradr.ai or have any questions? Contact us and we'll help you out.

For new inquiries:

For customer service:

Gradr.ai - Tribble AS

Bergstien 11B

0172 Oslo


NO 931992449 MVA


Last updated: 31.10.2026