VCR-Bench: leaderboard
Each model answers all 1,034 questions about 859 videos with step-by-step reasoning, and the score is the share of final answers GPT-4o grades correct (temporal localisation answers by IoU). Its sibling CoT Score board grades the reasoning steps rather than the final answer.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 9 models tracked. Updated 17 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Aria | 38.2 | #2189 |
| 2 | Qwen 2.5 VL 72B Instruct | 37.9 | #849 |
| 3 | LLaVA-Video-72B-Qwen2 | 36.6 | #1344 |
| 4 | InternVideo2.5-Chat-8B | 33 | #2031 |
| 5 | VideoLLaMA3-7B | 32.5 | #2015 |
| 6 | InternVL2.5-78B | 30.9 | #1036 |
| 7 | Qwen 2.5 VL 7B Instruct | 30.4 | #1893 |
| 8 | MiniCPM-o-2.6 | 26.9 | #2417 |
| 9 | InternVL2.5-8B | 23.9 | #2401 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=vcr-bench · How It Works · Data refreshed daily, snapshot 2026-11-17.