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

Language Model Alignment Evaluation on Arena-Hard v0.1

35.2Win Rate (%)

HyPO

4.20812.25420.328.346Feb 12, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.02
35.2
2026.02
33.5
2026.02
32.5
2026.02
32.1
2026.02
31.5
2026.02
31.3
2026.02
31
2026.02
30.5
2026.02
30.2
2026.02
30.1
2026.02
30
2026.02
28.6
2026.02
27.9
2026.02
25.2
2026.02
25.1
2026.02
23
2026.02
21
2026.02
20.5
2026.02
18.7
2026.02
18.4
2026.02
18.2
2026.02
18.2
2026.02
17.5
2026.02
16.9
2026.02
15.4
2026.02
14.3
2026.02
13.9
2026.02
12.2
2026.02
11.2
2026.02
9.5
2026.02
9.2
2026.02
8.2
2026.02
7.9
2026.02
6.6
2026.02
6.4
2026.02
5.4