AI BENCHMARK PROFILE

OmniRef-Bench

General AIMultimodal Perception

OmniRef-Bench is introduced in the paper but primarily serves to evaluate the proposed DyRef method. It is not presented as a standalone public benchmark.

Released: 2026-06-25
Readiness: Paper only
Primary field: General AI

Why it matters

N/A

Motivation

While personalized image generation has achieved remarkable progress, multi-reference image generation (MRIG) remains a challenging task.

Primary resources

Paper ↗Hugging Face ↗

Benchmark Radar records only publicly supported details and links back to primary sources for verification.