List your company

Video Reasoning on Maze (test)

82.1PrecisionVideoRLVR
10.54829.12447.766.276May 14, 2026

Evaluation results

MethodLinks
VideoRLVRModel Category=RL Model, Training Epochs=1 (from SFT Epoch 5)2026.0582.186.984.472.2
SFT Epoch 10Model Category=SFT Models, Training Epochs=102026.0580.485.182.769
SFT Epoch 5Model Category=SFT Models, Training Epochs=52026.0580.28381.666.1
VBVR-Wan2.2Model Category=SFT Models2026.0562.777.869.460.8
Kling V3Model Category=Proprietary2026.0524.815.719.223.5
Veo 3.1Model Category=Proprietary2026.0522.818.120.226
Wan-R1Model Category=SFT Models2026.0520.965.631.731.9
Wan2.2-TI2V-5BModel Category=Open-Source2026.0518.312.214.60
HunyuanVideoModel Category=Open-Source2026.0517.311.413.82.2
Sora 2Model Category=Proprietary2026.0515.817.216.53.1
CogVideoX1.5Model Category=Open-Source2026.0513.310.811.90