List your company
ResearchFormerly WizWand

Video Reasoning on VideoMMMU

68.33AccuracyVideo-Zero
46.458852.136957.81563.4931Oct 19, 2025Dec 1, 2025Jan 13, 2026Feb 26, 2026Apr 10, 2026May 23, 2026Jul 6, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Video-ZeroBackbone=MiMo-SFT-7B, Checkpoint=v32026.0568.33
Video-ZeroBackbone=MiMo-SFT-7B, Checkpoint=v42026.0567.78
Video-ZeroBackbone=MiMo-SFT-7B, Checkpoint=v12026.0567.56
Video-ZeroBackbone=MiMo-SFT-7B, Checkpoint=v52026.0567.56
Video-ZeroBackbone=MiMo-SFT-7B, Checkpoint=v22026.0567.44
MiMo-SFT-7BBackbone=MiMo-SFT-7B, Checkpoint=Base2026.0567.33
LATENT-VC-9BFrames=322026.0765.4
Video-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=52026.0565.33
Video-Zero (Peak)Backbone=Qwen3-VL-8B-Instruct2026.0565.33
LATENT-VC-9BFrames=642026.0765.3
LATENT-VC-9BFrames=162026.0765.2
V-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=12026.0564.33
VisPlayBackbone=Qwen3-VL-8B-Instruct, Iteration=52026.0564.22
V-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=22026.0564.22
Video-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=42026.0564.22
Video-Zero (Iter 4)Backbone=Qwen3-VL-8B-Instruct2026.0564.22
VisPlayBackbone=Qwen3-VL-8B-Instruct, Iteration=12026.0564.11
Base ModelBackbone=Qwen3-VL-8B-Instruct2026.0563.78
Video-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=12026.0563.78
Base ModelBackbone=Qwen3-VL-8B-Instruct2026.0563.78
VisPlayBackbone=Qwen3-VL-8B-Instruct, Iteration=32026.0563.67
V-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=32026.0563.67
Video-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=32026.0563.56
V-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=42026.0563.22
V-ZeroBackbone=Qwen3-VL-8B-Instruct2026.0563.22
Video-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=22026.0563.11
VisPlayBackbone=Qwen3-VL-8B-Instruct, Iteration=42026.0562.67
VisPlayBackbone=Qwen3-VL-8B-Instruct2026.0562.67
VisPlayBackbone=Qwen3-VL-8B-Instruct, Iteration=22026.0562.22
V-ZeroBackbone=Qwen3-VL-8B-Instruct, Iteration=52026.0562.22
Video-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=42026.0561.56
Video-Zero (Iter 4)Backbone=Qwen3-VL-4B-Instruct2026.0561.56
Video-Zero (Peak)Backbone=Qwen3-VL-4B-Instruct2026.0561.56
GPT-4o2025.1061.2
GPT-4oSize=-, Zero-shot=true, Category=Proprietary Models2026.0761.2
InternVL2.5Size=8B, Zero-shot=true, Category=Supervised Fine-Tuning Models2026.0760.8
Video-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=52026.0560.44
Qwen3.5-9BSFT+GRPOFrames=322026.0760.3
Qwen3.5-9BSFT+GRPOFrames=162026.0759.8
Qwen3.5-9BSFT+GRPOFrames=642026.0759.7
Video-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=32026.0559
Video-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=22026.0558.89
V-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=42026.0558.33
V-ZeroBackbone=Qwen3-VL-4B-Instruct2026.0558.33
VisPlayBackbone=Qwen3-VL-4B-Instruct, Iteration=12026.0557.89
V-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=52026.0557.89
V-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=12026.0557.67
V-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=32026.0557.67
Base ModelBackbone=Qwen3-VL-4B-Instruct2026.0557.56
Base ModelBackbone=Qwen3-VL-4B-Instruct2026.0557.56
VisPlayBackbone=Qwen3-VL-4B-Instruct, Iteration=42026.0557.44
VisPlayBackbone=Qwen3-VL-4B-Instruct2026.0557.44
VisPlayBackbone=Qwen3-VL-4B-Instruct, Iteration=32026.0557.22
VisPlayBackbone=Qwen3-VL-4B-Instruct, Iteration=52026.0557.11
Video-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=12026.0557.11
VisPlayBackbone=Qwen3-VL-4B-Instruct, Iteration=22026.0556.78
V-ZeroBackbone=Qwen3-VL-4B-Instruct, Iteration=22026.0556.33
Gemini-1.5-ProSize=-, Zero-shot=true, Category=Proprietary Models2026.0753.9
Video-ZeroBackbone=InternVL3.5-4B, Checkpoint=v12026.0553.11
TimeThinkSize=7B, Zero-shot=true, Category=Reinforcement Fine-Tuning Models2026.0752.6
Video-ZeroBackbone=InternVL3.5-4B, Checkpoint=v32026.0552.56
Video-ZeroBackbone=InternVL3.5-4B, Checkpoint=v52026.0552.56
VideoCFRFrames=322026.0652.4
Video-R1-7BFrames=642026.0752.4
Video-R1Size=7B, Zero-shot=true, Category=Reinforcement Fine-Tuning Models2026.0752.4
Video-R1-7BFrames=322026.0752.3
Video-ZeroBackbone=InternVL3.5-4B, Checkpoint=v42026.0552
VideoChat-R1Size=7B, Zero-shot=true, Category=Reinforcement Fine-Tuning Models2026.0752
Video-ZeroBackbone=InternVL3.5-4B, Checkpoint=v22026.0551.78
CAREFrames=642026.0651.1
VideoRFTSize=7B, Zero-shot=true, Category=Reinforcement Fine-Tuning Models2026.0751.1
CAREFrames=322026.0651
VideoCFRFrames=642026.0650.6
VideoCFRFrames=162026.0650.5
Qwen-2.5-VLSize=7B, Zero-shot=true, Category=Supervised Fine-Tuning Models2026.0750.3
Video-COM2026.0650.2
CAREFrames=162026.0650.2
Qwen2.5-VL-GRPOSize=7B, Zero-shot=true, Category=Reinforcement Fine-Tuning Models2026.0750.1
EvoVidModel=Qwen2.5-VL-7B-Instruct, Training Iteration=Iter 32026.0550
Video-R1-7BFrames=162026.0749.8
VideoChat-R1.5Size=7B, Zero-shot=true, Category=Reinforcement Fine-Tuning Models2026.0749.6
Qwen2.5-VL-SFTSize=7B, Zero-shot=true, Category=Supervised Fine-Tuning Models2026.0749.4
EvoVidModel=Qwen3-VL-8B-Instruct, Training Iteration=Iter 32026.0549.1
EvoVidModel=Qwen3-VL-8B-Instruct, Training Iteration=Iter 12026.0549
EvoVidModel=Qwen2.5-VL-7B-Instruct, Training Iteration=Iter 12026.0548.9
Video-R1-7BFrames/Resolution=32/1282025.1048.8
VideoChat-R1Frames=162026.0648.7
EvoVidModel=Qwen2.5-VL-7B-Instruct, Training Iteration=Iter 22026.0548.3
EvoVidModel=Qwen3-VL-8B-Instruct, Training Iteration=Iter 22026.0548.1
Base ModelModel=Qwen2.5-VL-7B-Instruct, Training Iteration=w/o training2026.0547.8
EvoVidModel=Qwen3-VL-8B-Instruct, Training Iteration=Frozen Questioner2026.0547.8
V-Reason-7BFrames/Resolution=16/2562025.1047.8
Video-R1-7BFrames/Resolution=16/2562025.1047.8
EvoVidModel=Qwen3-VL-4B-Instruct, Training Iteration=Iter 32026.0547.6
Qwen2.5-VL-7BFrames/Resolution=16/2562025.1047.6
V-Reason-7B (Lite)Frames/Resolution=16/2562025.1047.6
V-Reason-7B (Lite)Frames/Resolution=32/1282025.1047.4
Qwen2.5-VL-7B2026.0647.4
Qwen2.5-VL-7B2026.0647.4
EvoVidModel=Qwen3-VL-4B-Instruct, Training Iteration=Iter 12026.0547.3