EcoG-Bench (Video-Omni) - Referent Classification: leaderboard
Metric: Classification accuracy (%): share of referents for which the model picks the right option from the clip's 6-8 candidates, over EcoG-Bench's 811 egocentric clips (English and Chinese) with underspecified deictic instructions resolved by co-speech pointing, levels L1-L4 with one to four referents, temperature 0 and JSON output (unparsable answers score as wrong); the model ingests the raw video with its audio track; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Flash | 71.4 | #93 |
| 2 | Gemini 3 Pro | 63.9 | #77 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=ecog-bench-video-omni-referent-classification · How It Works · Data refreshed daily, snapshot 2026-10-11.