OpenVision 3: A Family of Unified Visual Encoder for Both Understanding and Generation
Open paper
Benchmark
Model
Rank
Results
image-generation-on-imagenet-256x256
OpenVision 3 (RAE)
–
FID: 1.89
visual-question-answering-vqa-on-seed-seedbench
OpenVision 3 (VAE + L/2)
–
Accuracy: 66.0
Rank counts only results with a code link.