LLM Stats (LVBench): leaderboard
Long video benchmark testing understanding of videos 30+ minutes, requiring temporal reasoning, event tracking, and narrative comprehension over extended content.
Metric: Score (%). Source: llm-stats.com. 31 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Argon | 91.7 | #8 |
| 2 | Gemini 3.8 Flash | 87.1 | #21 |
| 3 | Gemini 3.7 Flash | 85.4 | #23 |
| 4 | Qwen3.8 Max | 81.8 | |
| 5 | Seed 2.1 Pro | 78 | #67 |
| 6 | Seed 2.1 Turbo | 76.8 | #86 |
| 7 | Qwen3.8-Flash-Next | 76.6 | #72 |
| 8 | Qwen3.8 Flash | 76.6 | |
| 9 | Qwen3.7-Plus | 76.2 | |
| 10 | Kimi K2.5 | 75.9 | #144 |
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-lvbench · How It Works · Data refreshed daily, snapshot 2026-10-20.