Lab results
We give several AI tools the same real task, keep their raw output, and score it on stated criteria.
Every number here comes from an actual run. When something hasn't been tested yet, we say so.
Meeting transcription and notes
Turn a real recorded meeting into an accurate transcript and useful notes with decisions and action items.
Summarising a long PDF: are the facts right?
Summarise a 40-page report for a busy manager without dropping key numbers or inventing any.
Image generation for real business use
Make a logo, a product photo and a poster with exact text that a small business could actually use.
AI coding help for a non-coder building a small app
Someone who has never coded builds a small working app by following one AI tool's instructions.
Research with sources: accurate answers, real citations?
Answer research questions correctly and back every claim with a source that exists and says what the tool claims.
Runner self-check: summarising the US Constitution
Summarise a public-domain document with many numbers and keep every number right.