Video Reasoning on MMVU
75.8AccuracyGemini 2.5 Pro
Evaluation results
| Method | Links | |
|---|---|---|
| Gemini 2.5 ProEvaluation frame rate=1 FPS, self-collected via API=true2025.05 | 75.8 | |
| Seed1.5-VLThinking mode=true, Evaluation frame rate=1 FPS2025.05 | 70.1 | |
| Seed1.5-VLThinking mode=false, Evaluation frame rate=1 FPS2025.05 | 70.1 | |
| V-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=52026.05 | 68.96 | |
| V-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=42026.05 | 68.8 | |
| V-ZeroBackbone=Qwen3-VL-8B-Instruct2026.05 | 68.8 | |
| LATENT-VC-9BFrames=642026.07 | 68.6 | |
| FFR (GLM-4.5V)Distillation Method=FFR, Teacher Model=GLM-4.5V2026.04 | 68.5 | |
| V-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=22026.05 | 68.48 | |
| Base ModelBackbone=Qwen3-VL-8B-Instruct2026.05 | 68.32 | |
| Video-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=22026.05 | 68.32 | |
| Base ModelBackbone=Qwen3-VL-8B-Instruct2026.05 | 68.32 | |
| Video-Zero (Peak)Backbone=Qwen3-VL-8B-Instruct2026.05 | 68.32 | |
| FFRDistillation Method=FFR, Teacher Model=Qwen3-235B2026.04 | 68.2 | |
| VisPlayBackbone=Qwen3-VL-8B-Instruct, Iteration=12026.05 | 68.16 | |
| Video-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=52026.05 | 68.16 | |
| Video-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=42026.05 | 68 | |
| Video-Zero (Iter 4)Backbone=Qwen3-VL-8B-Instruct2026.05 | 68 | |
| FFRDistillation Method=FFR, Teacher Model=Qwen3-32B2026.04 | 67.9 | |
| LATENT-VC-9BFrames=322026.07 | 67.7 | |
| VisPlayBackbone=Qwen3-VL-8B-Instruct, Iteration=22026.05 | 67.68 | |
| VisPlayBackbone=Qwen3-VL-8B-Instruct, Iteration=32026.05 | 67.68 | |
| V-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=12026.05 | 67.52 | |
| SFTDistillation Method=SFT, Teacher Model=Qwen3-235B2026.04 | 67.4 | |
| VisPlayBackbone=Qwen3-VL-8B-Instruct, Iteration=52026.05 | 67.36 | |
| V-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=32026.05 | 67.36 | |
| Video-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=32026.05 | 67.36 | |
| Video-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=12026.05 | 67.2 | |
| Qwen3.5-9BSFT+GRPOFrames=642026.07 | 67.2 | |
| VisPlayBackbone=Qwen3-VL-8B-Instruct, Iteration=42026.05 | 67.04 | |
| VisPlayBackbone=Qwen3-VL-8B-Instruct2026.05 | 67.04 | |
| Qwen3.5-9BSFT+GRPOFrames=322026.07 | 66.6 | |
| LATENT-VC-9BFrames=162026.07 | 66.6 | |
| Video-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=22026.05 | 66.56 | |
| Video-Zero (Peak)Backbone=Qwen3-VL-4B-Instruct2026.05 | 66.56 | |
| Qwen3.5-9BSFT+GRPOFrames=162026.07 | 65.9 | |
| Video-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=52026.05 | 65.6 | |
| Video-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=42026.05 | 64.96 | |
| Video-Zero (Iter 4)Backbone=Qwen3-VL-4B-Instruct2026.05 | 64.96 | |
| VisPlayBackbone=Qwen3-VL-4B-Instruct, Iteration=52026.05 | 64.8 | |
| V-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=22026.05 | 64.8 | |
| V-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=42026.05 | 64.48 | |
| V-ZeroBackbone=Qwen3-VL-4B-Instruct2026.05 | 64.48 | |
| Base ModelBackbone=Qwen3-VL-4B-Instruct2026.05 | 64.32 | |
| VisPlayBackbone=Qwen3-VL-4B-Instruct, Iteration=12026.05 | 64.32 | |
| V-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=32026.05 | 64.32 | |
| Video-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=12026.05 | 64.32 | |
| Base ModelBackbone=Qwen3-VL-4B-Instruct2026.05 | 64.32 | |
| Video-R1-7BFrames=162026.07 | 64.2 | |
| V-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=12026.05 | 64.16 | |
| V-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=52026.05 | 64.16 | |
| SFTDistillation Method=SFT, Teacher Model=Qwen3-32B2026.04 | 63.9 | |
| Video-R1-7BFrames=322026.07 | 63.8 | |
| Video-R1-7BFrames=642026.07 | 63.8 | |
| VisPlayBackbone=Qwen3-VL-4B-Instruct, Iteration=22026.05 | 63.52 | |
| VisPlayBackbone=Qwen3-VL-4B-Instruct, Iteration=32026.05 | 63.36 | |
| VisPlayBackbone=Qwen3-VL-4B-Instruct, Iteration=42026.05 | 63.36 | |
| VisPlayBackbone=Qwen3-VL-4B-Instruct2026.05 | 63.36 | |
| Video-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=32026.05 | 63.2 | |
| Qwen3.5-9BCoTFrames=162026.07 | 61.6 | |
| Qwen3.5-9BCoTFrames=642026.07 | 61.6 | |
| Video-R1-SFTDistillation Method=SFT, Teacher Model=Baseline2026.04 | 61.3 | |
| Qwen3.5-9BCoTFrames=322026.07 | 59.8 | |
| Video-ZeroBackbone=InternVL3.5-4B, Checkpoint=v32026.05 | 58.88 | |
| Video-ZeroBackbone=InternVL3.5-4B, Checkpoint=v42026.05 | 58.72 | |
| Video-ZeroBackbone=InternVL3.5-4B, Checkpoint=v22026.05 | 58.4 | |
| Video-ZeroBackbone=InternVL3.5-4B, Checkpoint=v52026.05 | 58.4 | |
| Video-ZeroBackbone=InternVL3.5-4B, Checkpoint=v12026.05 | 57.6 | |
| InternVL3.5-4BCheckpoint=Base2026.05 | 54.08 | |
| LLaVA-OneVision-7BFrames=-2026.07 | 49.2 | |
| VideoLLaMA2Frames=-2026.07 | 44.8 |