Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
LLM Serving on MATH-plus (Average latency)
Loading...
107.1
Average Latency (ms/token)
Oracle SJF
99.45
151.0875
202.725
254.3625
Sep 25, 2025
Average Latency (ms/token)
Updated 26d ago
Evaluation Results
Method
Method
Links
Average Latency (ms/token)
Oracle SJF
Model=DeepSeek-R1
2025.09
107.1
PARS
Model=DeepSeek-R1
2025.09
115.65
Listwise SJF
Model=DeepSeek-R1
2025.09
123.69
Pointwise SJF
Model=DeepSeek-R1
2025.09
132.91
Cross-Model PARS
Model=DeepSeek-R1
2025.09
134.11
Oracle SJF
Model=Llama
2025.09
140.78
PARS
Model=Llama
2025.09
154.25
Cross-Model PARS
Model=Llama
2025.09
164.8
Listwise SJF
Model=Llama
2025.09
165.87
Pointwise SJF
Model=Llama
2025.09
167.76
FCFS
Model=DeepSeek-R1
2025.09
289.41
FCFS
Model=Llama
2025.09
298.35
Feedback
Search any
task
Search any
task