List your company

Question Answering on RAG QA 5 retrieved documents

191.7TTFT (ms)Lazy-Attn
149.248435.799722.351,008.901Jun 3, 2026

Evaluation results

MethodLinks
Lazy-AttnModel=Tulu3-Block-FT 8B2026.06191.76.2
Lazy-AttnModel=Llama-3.1-70B, Tensor Parallelism (TP)=42026.06238.45.2
CacheBlendModel=Tulu3-Block-FT 8B2026.06274.84.4
CacheBlendModel=Llama-3.1-70B, Tensor Parallelism (TP)=42026.06365.53.4
Standard RAGModel=Tulu3-Block-FT 8B2026.061,196.81
Standard RAGModel=Llama-3.1-70B, Tensor Parallelism (TP)=42026.061,2531