Skip to content

Leaderboard

Preference ratings from blind human votes, on an Elo-like scale — cost and speed were never visible to voters.

1Moonshot

Kimi K3

1605

±130

10/29 picked · $0.424/gen

provisional

2Anthropic

Claude Fable 5

1578

±132

9/30 picked · $1.04/gen

provisional

3OpenAI

GPT 5.6 Sol

1508

±144

6/30 picked · $0.305/gen

provisional

#ModelRatingVotesAvg cost
1
Kimi K3Moonshotprovisional
1605±130
29$0.424
2
Claude Fable 5Anthropicprovisional
1578±132
30$1.04
3
GPT 5.6 SolOpenAIprovisional
1508±144
30$0.305
4
GLM 5.2Z.aiprovisional
1502±143
31$0.066
5
Claude Opus 5Anthropicprovisional
1419±381
2$1.26

Ratings come from one preference model fitted to every blind vote at once (Plackett–Luce, on an Elo-like scale where +400 ≈ 10:1 odds). Because it’s fitted globally rather than accumulated vote by vote, a model added later is judged on its own record — not penalised for arriving late. ± is the 95% confidence interval; overlapping intervals mean the gap isn’t established yet. “Provisional” means too few votes to place confidently. W/L is times picked vs times shown. Cost and time are per full generation. ▲/▼ is the 7-day rank move.

Rating vs cost

Up and left is the value frontier — preferred by humans, cheap to run. The leader is highlighted in gold.

14001450150015501600$0.05$0.1$0.2$0.5$1.00$2.00Avg cost per generation (log scale)RatingKimi K3Claude Fable 5GPT 5.6 SolGLM 5.2Claude Opus 5

Bubble size reflects vote volume.