Preference Optimization
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
53Helpfulness Score
30
Jun 8, 2026
90.9Winrate
10
Apr 20, 2026
85.3Win Rate
5
Mar 4, 2026
76.97Win Rate
5
Mar 4, 2026
85.66Win Rate
5
Mar 4, 2026
0.8Avg Pref Score vs QWEN2.5-0.5B
5
Feb 26, 2026