LOL Arena (Earlier Runs): leaderboard
Metric: Human judge-choice agreement (%), earlier pipeline; denominator/settings unspecified. Source: goodstartlabs.com. Saturation forecast: Around 2033. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro | 49.05 |
| 2 | Grok 4 Fast | 47.1 |
| 3 | DeepSeek R1 | 46.1 |
| 4 | GPT-OSS-120B | 45.6 |
| 5 | GPT-5 Mini | 45.6 |
| 6 | GLM-4.6 | 45.6 |
| 7 | Llama 4 Maverick | 44.8 |
| 8 | Claude Haiku 4.5 | 44.6 |
| 9 | Hermes 4 405B | 44.6 |
| 10 | Kimi K2 0905 | 43.5 |
| 11 | AFM-4.5B | 40.5 |
Interactive version: theaggregate.ai/benchmark?slug=lol-arena-earlier-runs · How It Works · Data refreshed daily, snapshot 2026-10-09.