Video Understanding on VideoMME w/o sub
82.5ScoreQwen3.5
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3.5Model Parameters=35B-A3B, Subtitle Inclusion=false2026.06 | 82.5 | |
| Qwen3-VL ThinkingModel Parameters=235B-A22B, Thinking Mode=true, Subtitle Inclusion=false2026.06 | 79 | |
| GPT-5-miniRelease Date=2025-08-07, Subtitle Inclusion=false2026.06 | 78.9 | |
| Keye-VL-2.0-30B-A3BModel Parameters=30B-A3B, Subtitle Inclusion=false2026.06 | 78.3 | |
| Qwen3-VL ThinkingModel Parameters=32B, Thinking Mode=true, Subtitle Inclusion=false2026.06 | 77.3 | |
| Qwen3.5-OmniOpen-Source=false, Size=Flash2026.04 | 77 | |
| Gemini 2.5 FlashSize=-, mode=instruct2026.04 | 75.6 | |
| Gemini 1.5-Pro2026.03 | 75 | |
| Qwen3-VL ThinkingModel Parameters=30B-A3B, Thinking Mode=true, Subtitle Inclusion=false2026.06 | 73.3 | |
| InternVL3.5Model Parameters=241B-A28B, Subtitle Inclusion=false2026.06 | 72.9 | |
| Nemotron 3 Nano OmniOpen-Source=true, Size=30B-A3B, Reasoning=on2026.04 | 72.2 | |
| GPT-4o2026.03 | 71.9 | |
| Qwen3-VLSize=8B, mode=instruct2026.04 | 71.4 | |
| Nemotron 3 Nano OmniOpen-Source=true, Size=30B-A3B, Reasoning=off2026.04 | 70.8 | |
| Qwen3-OmniOpen-Source=true, Size=30B-A3B2026.04 | 70.5 | |
| Qwen3-OmniSize=30B-A3B, mode=instruct2026.04 | 70.5 | |
| MiniCPM-o 4.5Size=9B, mode=instruct2026.04 | 70.4 | |
| NVILA-8B-Video + AutoGazeMax #F=1024, Max Res.=35842026.03 | 67 | |
| VideoLLaMA3-7BMax #F=180, Max Res.=3842026.03 | 66.2 | |
| VideoLLaMA3-7B2025.05 | 66.2 | |
| InternVL3.5-8BMax #F=64, Max Res.=4482026.03 | 66 | |
| Nemotron Nano V2 VLOpen-Source=true, Size=12B, Reasoning=off2026.04 | 66 | |
| InternVL3.5Size=8B, mode=instruct2026.04 | 66 | |
| VideoChat-FlashMax #F=10000, Max Res.=4482026.03 | 65.3 | |
| LongVILA-R1-7BMax #F=8192, Max Res.=4482026.03 | 65.1 | |
| Qwen2.5-VL-7BMax #F=48, Max Res.=8962026.03 | 65.1 | |
| Qwen2.5 VL-7B2025.05 | 65.1 | |
| Gemini 2.5 Flash-Lite2026.03 | 65 | |
| NVILA-8B-VideoMax #F=256, Max Res.=4482026.03 | 64.2 | |
| Nemotron Nano V2 VLOpen-Source=true, Size=12B, Reasoning=on2026.04 | 63 | |
| Qwen2.5 VL-3B2025.05 | 61.5 | |
| Apollo-7BMax #F=2FPS, Max Res.=3842026.03 | 61.3 | |
| LongVILA-7BMax #F=2048, Max Res.=3842026.03 | 60.1 | |
| VideoLLaMA3-2B2025.05 | 59.3 | |
| VideoSafety-R1backbone=VideoLLaMA3-2B2025.05 | 59 | |
| LLaVA-OV-8BMax #F=32, Max Res.=3842026.03 | 58.2 |