Benchmark
VideoMMMU
Vision benchmarks measure how well models understand images. Scores are percentages. Higher is better.
- Models measured
- 27
- Scores recorded
- 38
- Sources
- 2
- IQ weight
- 3.33
Score distribution
- 82%
- Mean
- 84.4%
- Median
- 6.1%
- Std dev
- 74.7%
- 10th percentile
- 86.9%
- 90th percentile
- 61.2%
- Min
This benchmark has a weight of 3.33 for chat models, measured over 27 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Qwen3.8 MaxQwen3.8-Max release vision/video table · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 88.7% |
| 2 | Gemini 3 Pro Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 87.6% |
| 3 | Gemini 3 Pro PreviewGemini 3 Pro Model Card · results as of November 2025· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 87.6% |
| 4 | Gemini 3 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 86.9% |
| 5 | Gemini 3 Flashmedia_resolution: HIGH · temperature=0· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 86.9% |
| 6 | Kimi K2.5setting not named by the source· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 86.6% |
| 7 | Kimi K2.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 86.6% |
| 8 | GPT 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.9% |
| 9 | GPT 5.2no tools · xhigh· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 85.9% |
| 10 | Qwen3.7 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 85.4% |
| 11 | Qwen3.7 PlusQwen3.7-Plus release · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 85.4% |
| 12 | Gemini 3.1 Flash Lite· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84.8% |
| 13 | Gemini 3.1 Flash LiteGemini 3.1 Flash-Lite Model Card · high thinking · default sampling · single attempt/pass@1· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 84.8% |
| 14 | Qwen3.5 397B A17BQwen3.5 release/model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 84.7% |
| 15 | GPT 5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84.6% |
| 16 | MiniMax M3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84.6% |
| 17 | MiniMax M31 FPS · ≤512 frames · judge scoring· Reported by MiniMax Research· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 84.6% |
| 18 | Qwen3.6 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84.4% |
| 19 | Qwen3.6 27BQwen official release table · thinking mode unless benchmark requires otherwise· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 84.4% |
| 20 | Qwen3.6 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84% |
| 21 | Qwen3.6 PlusQwen official release table · thinking mode unless benchmark requires otherwise· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 84% |
| 22 | Qwen3.6 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.7% |
| 23 | Qwen3.6 FlashQwen3.6-35B-A3B official model card · hosted API name Qwen3.6-Flash· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 83.7% |
| 24 | Gemini 2.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.6% |
| 25 | OpenAI o3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 83.3% |
| 26 | Doubao Seed 1.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 82.7% |
| 27 | Qwen3.5 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 82.3% |
| 28 | Qwen3.5 122B-A10B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 82% |
| 29 | Seed 2.0 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 80.6% |
| 30 | Qwen3.5 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 80.4% |
| 31 | Qwen3 VL 235Bthinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 80% |
| 32 | Qwen3 VL 32B Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 79% |
| 33 | qwen3-vl-30bthinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 75% |
| 34 | Qwen3 VL 235B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 74.7% |
| 35 | Qwen: Qwen3 VL 8B Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 72.8% |
| 36 | qwen3-vl-30b· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 68.7% |
| 37 | Qwen3 VL 8B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. | 65.3% |
| 38 | GPT 4o· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 61.2% |
Data credits: Data by LLM Stats · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.