List your company

Mathematical Reasoning on Math500 (Throughput and Speedup)

20,417Throughput (tok/s)DFlash
-577.484,873.0110,323.515,773.99Feb 5, 2026

Evaluation results

MethodLinks
DFlashModel=Qwen3-4B, Concurrency=32, Inference Engine=SGLang, Backend=FA42026.0220,4172.98.01
DFlashModel=Qwen3-8B, Concurrency=32, Inference Engine=SGLang, Backend=FA42026.0216,0762.88.01
DFlashModel=Qwen3-4B, Concurrency=16, Inference Engine=SGLang, Backend=FA42026.0214,4773.58.01
DFlashModel=Qwen3-8B, Concurrency=16, Inference Engine=SGLang, Backend=FA42026.0212,2683.98.01
DFlashModel=Qwen3-4B, Concurrency=8, Inference Engine=SGLang, Backend=FA42026.029,0664.18.01
DFlashModel=Qwen3-8B, Concurrency=8, Inference Engine=SGLang, Backend=FA42026.027,4854.58.01
BaselineModel=Qwen3-4B, Concurrency=32, Inference Engine=SGLang, Backend=FA42026.027,136——
BaselineModel=Qwen3-8B, Concurrency=32, Inference Engine=SGLang, Backend=FA42026.025,694——
DFlashModel=Qwen3-4B, Concurrency=4, Inference Engine=SGLang, Backend=FA42026.024,9434.38.01
BaselineModel=Qwen3-4B, Concurrency=16, Inference Engine=SGLang, Backend=FA42026.024,100——
DFlashModel=Qwen3-8B, Concurrency=4, Inference Engine=SGLang, Backend=FA42026.023,8844.58.01
BaselineModel=Qwen3-8B, Concurrency=16, Inference Engine=SGLang, Backend=FA42026.023,133——
BaselineModel=Qwen3-4B, Concurrency=8, Inference Engine=SGLang, Backend=FA42026.022,201——
BaselineModel=Qwen3-8B, Concurrency=8, Inference Engine=SGLang, Backend=FA42026.021,666——
DFlashModel=Qwen3-4B, Concurrency=1, Inference Engine=SGLang, Backend=FA42026.021,5314.88.01
DFlashModel=Qwen3-8B, Concurrency=1, Inference Engine=SGLang, Backend=FA42026.021,1755.18.01
BaselineModel=Qwen3-4B, Concurrency=4, Inference Engine=SGLang, Backend=FA42026.021,145——
BaselineModel=Qwen3-8B, Concurrency=4, Inference Engine=SGLang, Backend=FA42026.02861——
BaselineModel=Qwen3-4B, Concurrency=1, Inference Engine=SGLang, Backend=FA42026.02316——
BaselineModel=Qwen3-8B, Concurrency=1, Inference Engine=SGLang, Backend=FA42026.02230——