Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
KV Cache Efficiency on Trace-driven Simulation (No-limit Budget, 100k Pool)
Loading...
21.33
Hit Ratio (Low Skewness)
LazyAttention
-0.25
5.3525
10.955
16.5575
Jun 3, 2026
Hit Ratio (Low Skewness)
Hit Ratio (Mid Skewness)
Hit Ratio (High Skewness)
Updated 1mo ago
Evaluation Results
Method
Method
Links
Hit Ratio (Low Skewness)
Hit Ratio (Mid Skewness)
Hit Ratio (High Skewness)
LazyAttention
KV Cache Mem Size=No-l...
2026.06
21.33
29.09
24.5
Block-Attention (vLLM)
KV Cache Mem Size=No-l...
2026.06
18.55
27.38
23.61
CacheBlend
KV Cache Mem Size=No-l...
2026.06
15.21
22.45
19.36
Prefix Caching
KV Cache Mem Size=No-l...
2026.06
0.58
2.16
3.25
Feedback
Search any
task
Search any
task