KV-cache compression

CacheFlow

CacheFlow: Efficient LLM Serving with 3D-Parallel KV Cache Restoration

On the frontierfirst seen Apr 28, 2026

Current frontier — recent, not yet superseded in the knowledge base

0 papers critique it · 0 beat it on benchmarks

What to use instead

Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.