LLM Stats (MLVU-M): leaderboard
Multi-task Long Video Understanding (Multiple choice): multiple-choice variant testing long video comprehension with structured answer selection.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Estimated already saturated. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 VL 32B Instruct | 82.1 | #283 |
| 2 | Qwen 3 VL 30B A3B Instruct | 81.3 | #384 |
| 3 | Qwen 3 VL 30B A3B (Thinking) | 78.9 | #351 (Qwen 3 VL 30B A3B) |
| 4 | Qwen 3 VL 8B Instruct | 78.1 | #420 |
| 5 | Qwen 3 VL 4B (Thinking) | 75.7 | #480 (Qwen 3 VL 4B) |
| 6 | Qwen 3 VL 4B Instruct | 75.3 | #515 |
| 7 | Qwen 3 VL 8B (Thinking) | 75.1 | |
| 8 | Qwen 2.5 VL 72B Instruct | 74.6 | #377 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-mlvu-m · How It Works · Data refreshed daily, snapshot 2026-10-20.