WorldCupBench - Brier Skill: leaderboard
WorldCupBench live tournament forecasting score transformed as 100 minus total Brier score, so higher is better.
Metric: 100 - Brier Total. Source: github.com. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V4 Pro | 85.35 | #102 |
| 2 | MiMo-V2.5-Pro | 85.35 | #158 |
| 3 | MiniMax-M3 | 85.32 | #136 |
| 4 | Nex N2 Pro | 85.17 | #107 |
| 5 | Gemini 3.5 Flash | 85.14 | #60 |
| 6 | GPT-5.5 | 85.03 | #31 |
| 7 | Qwen 3.7 Max | 84.94 | #75 |
| 8 | Grok 4.3 | 84.92 | #188 |
| 9 | Kimi K2.6 | 84.86 | #104 |
| 10 | Claude Fable 5 | 84.78 | #16 |
| 11 | GLM-5.1 | 84.61 | #97 |
Interactive version: theaggregate.ai/benchmark?slug=worldcupbench-brier-skill · How It Works · Data refreshed daily, snapshot 2026-10-20.