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

Multi-turn Conversation Evaluation on MT-bench (Speedup, tau)

3.82Speedup

Eagle3+LTD

1.46962.07982.693.3002Mar 2, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.03
3.826.4
2026.03
3.615.58
2026.03
3.65.55
2026.03
3.586.3
2026.03
3.485.55
2026.03
3.435.36
2026.03
3.45.05
2026.03
3.385.66
2026.03
3.185.73
2026.03
3.185.2
2026.03
3.074.74
2026.03
3.055.19
2026.03
2.924.2
2026.03
2.893.4
2026.03
2.854.11
2026.03
2.854.02
2026.03
2.823.82
2026.03
2.794.08
2026.03
2.784.11
2026.03
2.753.54
2026.03
2.743.41
2026.03
2.172.78
2026.03
2.163.11
2026.03
2.053.13
2026.03
1.562.89