KV cache reuse efficiency on LMSys
47.3Match RateCachePrune
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CachePrunePrivacy setting=disabled, Reuse granularity=variable-length fine-grained, Model=Mistral-7B2026.05 | 47.3 | 74 | |
| CacheBlendReuse granularity=fixed-size chunks, Chunk size=128 tokens, Model=Mistral-7B2026.05 | 3.1 | 93 | |
| EPICReuse granularity=fixed-size chunks, Chunk size=128 tokens, Model=Mistral-7B2026.05 | 3.1 | 94 | |
| CacheCraftReuse granularity=fixed-size chunks, Chunk size=128 tokens, Model=Mistral-7B2026.05 | 2.9 | 98 |