The best AI model is .
Frontier models get the same brief. You pick the best, blind. Then the unmasking — who made what, what it cost, how fast it was.
blind votes 45 · comparisons 172 · models 5 · challenges 15
The protocol
- 01
Inspect
Open a challenge and inspect the outputs side by side — no names, no logos, no prices.
- 02
Pick one
Crown the best of them. Every pick feeds one global preference model.
- 03
Unmask
The bars lift: who made what, what it cost, how long it took.
Current standings
| # | Model | Rating | Votes | Avg cost |
|---|---|---|---|---|
| 1 | Claude Opus 5anthropic | 1591 | 6 | $1.26 |
| 2 | Claude Fable 5anthropic | 1563 | 64 | $1.04 |
| 3 | GPT 5.6 Solopenai | 1496 | 76 | $0.305 |
| + 2 more models on the leaderboard | ||||
Ratings move with every blind vote. Cost was never visible to voters.