List your company

Visual Equivalence Reward Modeling on Visual-ERM-Bench Chart

39.9F1hVisual-ERM
1.00411.10221.231.298Mar 13, 2026

Evaluation results

MethodLinks
Visual-ERMCategory=Open-source, Backbone=Qwen3-VL-8B-Instruct2026.0339.942.861.2
Gemini-3-FlashCategory=Proprietary2026.0338.541.362.8
Gemini-2.5-ProCategory=Proprietary2026.0333.737.561.8
GPT-5.2Category=Proprietary2026.0330.132.664.8
Qwen3-VL-235B-InstructCategory=Open-source2026.032831.847.2
GPT-4oCategory=Proprietary2026.0322.828.348.5
Qwen2.5-VL-7BCategory=Open-source2026.033.95.411.2
Qwen3-VL-8B-InstructCategory=Open-source2026.033.33.53.8
InternVL3.5-8BCategory=Open-source2026.032.55.711