KV-cache compression

Global Retention-Based KV Eviction

Make Each Token Count: Towards Improving Long-Context Performance with KV Cache Eviction

On the frontierfirst seen May 10, 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Current frontier — recent, not yet superseded in the knowledge base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Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.