Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
KV Cache Efficiency on Trace-driven Simulation (1GB Budget, 100k Pool)
Loading...
3.47
Low Skewness Hit Ratio
LazyAttention
-0.1388
0.7981
1.735
2.6719
Jun 3, 2026
Low Skewness Hit Ratio
Mid Skewness Hit Ratio
High Skewness Hit Ratio
Updated 1mo ago
Evaluation Results
Method
Method
Links
Low Skewness Hit Ratio
Mid Skewness Hit Ratio
High Skewness Hit Ratio
LazyAttention
KV Cache Mem Size=1 GB...
2026.06
3.47
11.11
13.57
Block-Attention (vLLM)
KV Cache Mem Size=1 GB...
2026.06
1.84
6.03
7.27
CacheBlend
KV Cache Mem Size=1 GB...
2026.06
1.51
4.95
5.96
Prefix Caching
KV Cache Mem Size=1 GB...
2026.06
0
0
0
Feedback
Search any
task
Search any
task