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

Multi-turn Conversation Evaluation on MT-bench

81.1MT-Bench Score

+RL (Skywork-Reward-V2-Llama-3.1-8B)

-3.0474418.7985340.644562.49047May 30, 2025Jul 26, 2025Sep 22, 2025Nov 18, 2025Jan 15, 2026Mar 13, 2026May 10, 2026
Updated 22d ago

Evaluation Results

MethodLinks
81.1--
2025.07
79--
2025.07
78.8--
78.2--
76.8--
2025.07
67.3--
2025.07
66.5--
2025.07
65.7--
2025.07
63.5--
2025.07
62.8--
2025.07
58.5--
2025.07
57.1--
2025.07
56.8--
2025.07
52.8--
2026.05
8.79--
2026.05
8.61--
2026.05
8.56--
2026.05
8.55--
2026.05
8.54--
2026.05
8.18--
2026.05
8.13--
2026.05
8.1--
2026.05
8.07--
2026.05
8.04--
2026.05
8.02--
2026.05
8.01--
2026.05
8.01--
2026.05
7.92--
2026.05
7.63--
2025.06
7.62--
2026.05
7.62--
2026.05
7.59--
2026.05
7.58--
2026.05
7.57--
2026.05
7.54--
2026.05
7.53--
2026.05
7.52--
2026.05
7.51--
2026.05
7.51--
2026.05
7.5--
2026.05
7.47--
2026.05
7.43--
2026.05
7.4--
2026.05
7.39--
2025.06
7.34--
2026.05
7.29--
2025.06
7.21--
2025.05
7.07--
2025.05
7.07--
2026.05
7.06--
2026.05
6.92--
2025.05
6.89--
2026.05
6.86--
2025.06
6.42--
2025.05
6.37--
2025.06
6.29--
2025.06
6.16--
2026.05
5.4--
2026.05
0.794--
2026.05
0.763--
2026.05
0.734--
2026.05
0.686--
2026.05
0.632--
2026.05
0.618--
2026.05
0.606--
2026.05
0.555--
2026.05
0.467--
2026.05
0.189--
2024.09
-2.342.79
2024.09
-2.922.55
2024.09
-3.132.77
2024.09
-1.561.29
2024.09
-3.142.71
2024.09
-4.383.99
2024.09
-3.073.73
2024.09
-4.213.93
2024.09
-3.513.26
2024.09
-4.143.64