OmniGameArena - ObstacleRun3D — leaderboard

Metric: Normalized game score (%). Source: arxiv.org. 10 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.620
2Claude Opus 4.617.2
3Gemini 3.1 Pro (Preview)16.5
4GPT-5.513.3
5Qwen 3.5 397B A17B11.2
6Gemini 3.1 Flash Lite (Preview)9.7
7Claude Opus 4.79.4
8GPT-5.48.9
9Kimi K2.57.5
10Qwen 3.5 122B A10B3.4

Interactive version: aibenchmarks.dev/benchmark?slug=omnigamearena-obstaclerun3d · How the rankings work · Data refreshed daily, snapshot 2026-07-06.