CrossVid - Comparative Analysis: leaderboard
Mean accuracy over CrossVid's four comparative tasks (behaviour, narrative, culinary and procedural-error comparison across videos).
Metric: Accuracy (%). Source: raw.githubusercontent.com. Saturation forecast: Estimated already saturated. 16 models tracked. Updated 17 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 54.7 | #274 |
| 2 | GPT-4.1 | 47.6 | #505 |
| 3 | GLM-4.1V-9B (Thinking) | 44.7 | |
| 4 | Qwen 2.5 VL 32B Instruct | 38.3 | #1074 |
| 5 | LLaVA-Video-72B-Qwen2 | 33.9 | #1344 |
| 6 | Kimi-VL-A3B (Thinking) | 33.4 | |
| 7 | InternVL3-78B | 33.1 | #755 |
| 8 | InternVL3-38B | 27.8 | #927 |
| 9 | Qwen2.5-Omni-7B | 26.7 | #1883 |
| 10 | MiniCPM-o-2.6 | 26.2 | #2417 |
| 11 | InternVL3-8B | 26.1 | #1756 |
| 12 | Phi-3.5-Vision Instruct | 25.9 | #6271 |
| 13 | LongVA-7B-DPO | 23.5 | #2177 |
| 14 | VideoLLaMA3-7B | 20.8 | #2015 |
| 15 | Qwen 2.5 VL 7B Instruct | 19.3 | #1893 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=crossvid-comparative-analysis · How It Works · Data refreshed daily, snapshot 2026-11-17.