Our set-up
- We test with ordinary accounts, the same way you would, and note whether a feature needs a paid plan.
- We use exactly the same prompt, word for word, for every tool in a comparison.
- We run each test at least twice, because AI answers can change from one try to the next.
- We record the date we tested, because tools are updated often.
The six tests
- Accuracy: questions with answers we have already checked.
- Writing: a short email or post in a set tone and length.
- Reasoning: simple logic and maths problems with one right answer.
- Fresh knowledge: questions about recent events, to see whether the tool can find up-to-date information.
- Files: summarising a document or image we upload.
- Privacy: what the tool does with your data, and the controls it gives you.
How we score
Each test gets a score from 1 to 5. The total out of 30 decides our overall ranking, and we also note price, usage limits and ease of use.
The best tool for you depends on what you need, so each comparison also says which tool suits which kind of person.
When we retest
We retest a comparison when one of the tools has a major update, and at least every six months. The date of the latest test appears at the top of each comparison.
What our tests cannot tell you
Our tests are a snapshot of a fast-moving field. Results can vary with your account, region and the exact wording you use, so we encourage you to try a tool yourself before you pay for it.