List your company

Visual Equivalence Reward Modeling on Visual-ERM-Bench SVG

33.3F1h ScoreGemini-3-Flash
0.6449.12217.626.078Mar 13, 2026

Evaluation results

MethodLinks
Gemini-3-FlashCategory=Proprietary2026.0333.367.564.3
Gemini-2.5-ProCategory=Proprietary2026.0329.334.363.3
GPT-5.2Category=Proprietary2026.0328.532.261.1
Visual-ERMCategory=Open-source, Backbone=Qwen3-VL-8B-Instruct2026.0328.332.659.6
Qwen3-VL-235B-InstructCategory=Open-source2026.0319.422.851.5
GPT-4oCategory=Proprietary2026.031319.350.3
InternVL3.5-8BCategory=Open-source2026.036.113.148.9
Qwen3-VL-8B-InstructCategory=Open-source2026.036.19.427.1
Qwen2.5-VL-7BCategory=Open-source2026.031.97.537.9