EvaOmni - Recognition and Grounding - Answer Accuracy: leaderboard
EvaOmni (2026):1027questions from392videos,1FPS and480px long side with audio; shared prompt. Open non-MiMo models use recommended card configurations on8H20GPUs; MiMo and closed models use official APIs. Coverage/bottleneck grounding score correct answers against annotated audio and visual intervals; expert reference uses replay and is retained separately. Score (%).
Metric: Score (%). Source: arxiv.org. Saturation forecast: Around January 2027. 24 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.7 Flash | 85.3 | #22 |
| 2 | Gemini 3.8 Flash | 83.4 | #24 |
| 3 | Gemini 3.6 Flash | 80.9 | #51 |
| 4 | Gemini 2.5 Pro | 76.5 | #154 |
| 5 | Gemini 3.1 Pro (Preview) | 74.7 | #60 |
| 6 | Gemini 4 Argon | 67.1 | #8 |
| 7 | MiMo-V2.5 | 65.4 | #197 |
| 8 | Qwen3.8 Omni Flash | 65 | #33 |
| 9 | Gemini 3.5 Flash | 64.1 | #61 |
| 10 | MiMo-V2.6-Flash | 61.3 | #73 |
| 11 | Gemini 3.5 Flash Lite | 60.6 | #159 |
| 12 | MiMo-V2.6-Pro | 60.1 | #38 |
| 13 | Gemini 3.1 Flash Lite (Preview) | 59.7 | #198 |
| 14 | Gemini 3.1 Flash Lite | 59.4 | #194 |
| 15 | Gemini 2.5 Flash | 59.2 | #246 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=evaomni-recognition-and-grounding-answer-accuracy · How It Works · Data refreshed daily, snapshot 2026-10-22.