List your company

End-to-end single-step decoding on DeepSeek-R1-Distill-LLaMA-8B 32K Context

23.4Latency (ms)LessIsMore
23.224.5525.927.25Aug 9, 2025

Evaluation results

MethodLinks
LessIsMoreToken Budget=2K, Serving Stack=SGLang + FlashInfer, Hardware=NVIDIA A5000 GPU2025.0823.4
QuestToken Budget=2K, Serving Stack=SGLang + FlashInfer, Hardware=NVIDIA A5000 GPU2025.0824.4
TidalDecodeToken Budget=2K, Serving Stack=SGLang + FlashInfer, Hardware=NVIDIA A5000 GPU2025.0824.7
Full AttentionToken Budget=2K, Serving Stack=SGLang + FlashInfer, Hardware=NVIDIA A5000 GPU2025.0828.4