List your company

LLM Serving Performance on Qwen 7B Uniform distribution 32 model variants 2.5

0.86Throughput (req/s)Delta-CoMe
0.03840.25170.4650.6783Jun 5, 2025

Evaluation results

MethodLinks
Delta-CoMearrival rate (lambda)=1.02025.060.861.44
PRINMIXarrival rate (lambda)=1.02025.060.861.17
Delta-CoMearrival rate (lambda)=0.52025.060.420.81
PRINMIXarrival rate (lambda)=0.52025.060.420.79
Naivearrival rate (lambda)=1.02025.060.08481.42
Naivearrival rate (lambda)=0.52025.060.07253.93