C$^2$KV: Compressed and Composable KV Cache Reuse for Efficient LLM Inference

Researchers propose C$^2$KV, a unified framework for non-prefix KV cache reuse that jointly optimizes KV extraction and inference-time concatenation, achieving up to 17× inference speedup under long contexts while preserving generation quality. The method learns a composable and compressed KV cache manifold that is position-agnostic, using a lightweight sidecar Extractor with learnable compression tokens and structured attention flow to enable modular KV representations without modifying the frozen base model.

arXiv:2607.17715v1 Announce Type: new Abstract: Long-context inference is central to modern large language model LLM applications such as retrieval-augmented generation and multi-document reasoning. To mitigate the growing inference cost, recent work has explored key-value KV cache reuse to reduce redundant prefill computation. However, existing reuse methods primarily focus on computation savings and overlook a critical bottleneck in long-context LLM serving: the cost of storing and accessing large KV caches. While KV compression appears to be a natural complement, naively combining compression with non-prefix KV reuse often leads to severe accuracy degradation. In this work, we propose C$^2$KV, a unified framework for non-prefix KV reuse that jointly optimizes KV extraction and inference-time concatenation. C$^2$KV learns a composable and compressed KV cache manifold that is explicitly designed to be position-agnostic. Our approach introduces a lightweight sidecar Extractor with learnable compression tokens and a structured attention flow, enabling modular KV representations that can be flexibly reused and concatenated without modifying the frozen base model. We further employ a compression-concatenation co-training strategy to align extraction-time representations with their downstream reuse behavior. Extensive experiments across multiple long-context benchmarks and model families demonstrate that C$^2$KV significantly reduces KV cache storage and transfer costs, achieving up to 17$\times$ inference speedup under long contexts, while preserving generation quality.