Benchmark
LVBench
What is LVBench?
LVBench is an extreme long video understanding benchmark designed to evaluate multimodal models on videos up to two hours in duration. It contains 6 major…
- Released
- 2024-06-12
- Modality
- multimodal
- Evaluates
- long context, multimodal, vision
- Openness
- unknown
- Importer
- LLM Stats
- Reported scores
- 25
Source provenance
- Original evidence https://api.zeroeval.com/leaderboard/benchmarks/lvbench?top_n=500
LVBench code and dataset
LVBench results and reported scores
LLM Stats
| Model | Organization | Reported value | Reported | Evidence |
|---|---|---|---|---|
| Gemini 3.7 Flash | 0.854 | 2026-08-13 | evidence | |
| Kimi K2.5 | Moonshot AI | 0.759 | 2026-01-27 | evidence |
| Nova Lite | Amazon | 0.404 | 2024-11-20 | evidence |
| Nova Pro | Amazon | 0.416 | 2024-11-20 | evidence |
| Qwen2.5 VL 32B Instruct | Qwen | 0.49 | 2025-02-28 | evidence |
| Qwen2.5 VL 72B Instruct | Qwen | 0.473 | 2025-01-26 | evidence |
| Qwen2.5 VL 7B Instruct | Qwen | 0.453 | 2025-01-26 | evidence |
| Qwen3 VL 235B A22B Instruct | Qwen | 0.677 | 2025-09-22 | evidence |
| Qwen3 VL 235B A22B Thinking | Qwen | 0.636 | 2025-09-22 | evidence |
| Qwen3 VL 30B A3B Instruct | Qwen | 0.625 | 2025-09-22 | evidence |
| Qwen3 VL 30B A3B Thinking | Qwen | 0.592 | 2025-09-22 | evidence |
| Qwen3 VL 32B Instruct | Qwen | 0.638 | 2025-09-22 | evidence |
| Qwen3 VL 32B Thinking | Qwen | 0.626 | 2025-09-22 | evidence |
| Qwen3 VL 4B Instruct | Qwen | 0.562 | 2025-09-22 | evidence |
| Qwen3 VL 4B Thinking | Qwen | 0.535 | 2025-09-22 | evidence |
| Qwen3 VL 8B Instruct | Qwen | 0.58 | 2025-09-22 | evidence |
| Qwen3 VL 8B Thinking | Qwen | 0.558 | 2025-09-22 | evidence |
| Qwen3.5-122B-A10B | Qwen | 0.744 | 2026-02-24 | evidence |
| Qwen3.5-27B | Qwen | 0.736 | 2026-02-24 | evidence |
| Qwen3.5-35B-A3B | Qwen | 0.714 | 2026-02-24 | evidence |
| Qwen3.6-35B-A3B | Qwen | 0.714 | 2026-04-16 | evidence |
| Qwen3.7-Plus | Qwen | 0.762 | 2026-05-31 | evidence |
| Qwen3.8 Max | Qwen | 0.818 | 2026-08-02 | evidence |
| Seed 2.1 Pro | ByteDance | 0.78 | 2026-06-24 | evidence |
| Seed 2.1 Turbo | ByteDance | 0.768 | 2026-06-24 | evidence |
Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.