List your company

Speculative Decoding on Code (Throughput)

138.72Throughput (tokens/s)EvoSpec
101.9248111.4774121.03130.5826Apr 17, 2026

Evaluation results

MethodLinks
EvoSpecStatic Vocabulary Size=32k, Dynamic Buffer Size (Dyna)=2562026.04138.72
EvoSpecStatic Vocabulary Size=32k, Dynamic Buffer Size (Dyna)=5122026.04137.9
EvoSpecStatic Vocabulary Size=32k, Dynamic Buffer Size (Dyna)=10242026.04137.14
EvoSpecStatic Vocabulary Size=16k, Dynamic Buffer Size (Dyna)=10242026.04131.19
EvoSpecStatic Vocabulary Size=16k, Dynamic Buffer Size (Dyna)=5122026.04129.29
EvoSpecStatic Vocabulary Size=16k, Dynamic Buffer Size (Dyna)=2562026.04125.77
FR-SpecStatic Vocabulary Size=32k2026.04123.61
Vanilla SD2026.04119.56
EvoSpecStatic Vocabulary Size=16k, Dynamic Vocabulary Size=256, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04118
EvoSpecStatic Vocabulary Size=32k, Dynamic Vocabulary Size=256, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04116.38
EvoSpecStatic Vocabulary Size=16k, Dynamic Vocabulary Size=1024, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04116.3
EvoSpecStatic Vocabulary Size=16k, Dynamic Vocabulary Size=512, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04115.64
EvoSpecStatic Vocabulary Size=8k, Dynamic Buffer Size (Dyna)=10242026.04115.14
Standard SDDraft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04114.71
EvoSpecStatic Vocabulary Size=8k, Dynamic Vocabulary Size=256, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04114.07
EvoSpecStatic Vocabulary Size=8k, Dynamic Buffer Size (Dyna)=5122026.04113.82
EvoSpecStatic Vocabulary Size=32k, Dynamic Vocabulary Size=512, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04113.62
EvoSpecStatic Vocabulary Size=32k, Dynamic Vocabulary Size=1024, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04113.38
EvoSpecStatic Vocabulary Size=8k, Dynamic Vocabulary Size=512, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04113.34
EvoSpecStatic Vocabulary Size=8k, Dynamic Vocabulary Size=1024, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04112.63
EvoSpecStatic Vocabulary Size=8k, Dynamic Buffer Size (Dyna)=2562026.04112.51
FR-SpecStatic Vocabulary Size=32k, Draft Model=Qwen3-1.7B, Target Model=Qwen3-4B, Architecture=Decoupled2026.04103.34