Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Instruction-guided image editing preference prediction on GenAI-Bench

67.5Accuracy

REWARDHARNESS (Qwen)

24.23635.46846.757.932Sep 30, 2025Nov 5, 2025Dec 12, 2025Jan 18, 2026Feb 24, 2026Apr 2, 2026May 9, 2026
Updated 22d ago

Evaluation Results

MethodLinks
2026.05
67.5
2025.09
65.72
2026.05
65.7
2026.05
64.4
2026.05
64
2025.09
63.97
2025.09
59.96
2025.09
59.61
2026.05
59.6
2026.05
57.9
2025.09
57.89
2025.09
57.01
2026.05
57
2025.09
53.54
2026.05
53.5
2025.09
53.32
2026.05
53.3
2026.05
47.1
2025.09
42.76
2026.05
40.5
2025.09
40.48
2026.05
39.3
2025.09
39.28
2025.09
25.9