aiintelligence

The AI proving ground

We test AIso you don'thave to.

30 tools sit identical tests, marked blind by judges from rival labs. Our score beside what millions of real users say — updated daily.

All the verdictsNew to AI? Start here

every score has downloadable receipts · no vendor pays for placement

Live board — Everyday chat10 Aug 2026
1ChatGPT
90/100
2Claude
87/100
3Gemini
84/100the full everyday chat ranking, with receipts →

Best for…

One verdict per job, from tests we ran ourselves. Chat, writing and images are jobs anyone can use in a browser; research and meetings are still being tested.

Everyday chatChatGPTclose race
90/100
WritingChatGPTclose race
90/100
ImagesGPT Image (in ChatGPT) winsfrom head-to-head
battle →
ResearchIn testing
—soon
Meetings & notesIn testing
—soon
CodingGitHub Copilotclose race
90/100
Best freeGemini 3.5 Flash winsfrom head-to-head
battle →
Best-value APIGrok 4.5 winsfrom head-to-head
battle →

What changed this week

everything →
  1. www.notion.com/pricing content changed per Wayback snapshot 20260808102404 — price extractors will verify on the next daily scan10 Aug 2026
  2. cursor.com/pricing content changed per Wayback snapshot 20260808203757 — price extractors will verify on the next daily scan10 Aug 2026
  3. claude.com/pricing content changed per Wayback snapshot 20260809175743 — price extractors will verify on the next daily scan10 Aug 2026
How the testing worksSame tasks, blind three-lab judging, receipts on every score.the protocol →If you write codeModel battles, API prices per million tokens, leaderboards.for developers →

data last verified 10 Aug 2026 · no vendor pays for placement