List your company

Video Editing Evaluation on VEFX-Dataset (test)

0.78SRCC (Overall)VEFX-Reward-32B
0.191360.344180.4970.64982Apr 17, 2026

Evaluation results

MethodLinks
VEFX-Reward-32BMethod Category=Ours2026.040.780.7540.6810.7170.6120.5670.5970.6160.7510.7920.7320.790.8250.6430.740.475
VEFX-Reward-4BMethod Category=Ours2026.040.760.7140.690.6930.5640.5740.5560.5950.7040.7930.710.7710.8880.6420.7640.493
Gemini-3.1-ProMethod Category=VLM-as-a-Judge2026.040.7520.7310.5180.6810.5590.4590.5840.6080.7540.510.6440.7260.8260.8640.7880.546
Seed-2.0-LiteMethod Category=VLM-as-a-Judge2026.040.720.5450.5440.6970.4830.4970.5270.6070.6160.5940.7290.7680.9840.8150.7910.51
Qwen3.5-122BMethod Category=VLM-as-a-Judge2026.040.6310.3790.5630.6580.3270.5230.5730.520.3780.6630.6010.6851.1650.7520.8930.578
Seed-1.6Method Category=VLM-as-a-Judge2026.040.630.6860.6180.5040.6050.5730.4470.5080.6840.6080.5910.7010.9180.7980.9170.565
Qwen3.5-397BMethod Category=VLM-as-a-Judge2026.040.6010.5720.4220.6540.5060.3840.5870.4970.6150.6240.6920.6570.9920.7850.820.598
Gemini-3.1-Flash-LiteMethod Category=VLM-as-a-Judge2026.040.5740.3090.3020.6610.2830.2770.5550.4360.3160.4250.6730.5051.1940.910.840.685
EditRewardMethod Category=Previous Reward Models2026.040.5580.453-0.211—0.342-0.164—0.4110.4550.317—0.581.1130.844—0.631
Gemini-2.5-FlashMethod Category=VLM-as-a-Judge2026.040.3830.2560.2170.5810.2360.1950.5440.2960.2560.4910.5690.4781.2160.8750.9340.697
VE-BenchMethod Category=Previous Reward Models2026.040.214——————0.15———0.238———0.752