EgoEsportsQA - Reasoning: leaderboard
EgoEsportsQA (2026) is a video question-answering benchmark of 1,745 expert-checked multiple-choice questions on first-person clips from professional Counter-Strike 2, Valorant and Overwatch 2 matches. This board is accuracy in percent on the reasoning-level questions; higher is better.
Metric: Accuracy (%) on the reasoning-level EgoEsportsQA questions over the three games; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 9 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 67.98 | #178 |
| 2 | Seed 1.8 | 55.91 | #252 |
| 3 | Claude Sonnet 4.5 | 47.24 | #261 |
| 4 | Qwen 3 VL 8B Instruct | 43.44 | #1029 |
| 5 | Gemini 3 Flash | 40.81 | #175 |
| 6 | EgoGPT-7B | 35.43 | #2397 |
| 7 | InternVL3.5-8B | 35.17 | #1604 |
| 8 | LLaVA-OneVision-7B | 32.68 | #2981 |
| 9 | LLaVA-NeXT-Video-7B | 27.17 | #5734 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=egoesportsqa-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-21.