List your company

Visual Consistency Assessment on EditReward-Bench 2025 (test)

80.88Subject Addition AccuracyGemini 3 Pro
56.720862.992969.26575.5371Mar 30, 2026

Evaluation results

MethodLinks
Gemini 3 Pro2026.0380.8883.5690.5791.390.1688.8986.679278.9594.1284.9387.8485.4887.33
GPT 5.12026.0380.8863.0177.3686.9688.5287.583.338350.8889.7175.3475.6880.6578.68
PVC-Judge2026.0377.9467.1284.9194.283.6186.11808771.9389.7187.6782.4379.0382.44
EditReward2026.0374.3889.981.5583.496774.3972.6177.1593.6669.0284.3875.837578.34
Qwen3-VL-8B-Instruct2026.0372.0661.6473.5872.4675.4176.3978.337456.1485.2965.7574.3274.1972.27
EditScore Avg@42026.0357.6551.9154.8577.8351.2361.3855.6563.2576.0629.4151.5658.7562.2657.83