mlsanityData to 14 Oct 2026

KVCMAS: Efficient KV cache Correction for Shared Context in Multi-Agent Systems

arXiv 2609.34060 ↗First posted 28 Sep 2026cs.LGcited by 0

On the map: Efficient LLM inference and serving → Efficient reasoning and agent inference

Abstract (arXiv)

Prompt-specialized multi-agent systems enable multiple agents to share a model while performing complementary roles to solve complex tasks. However, agent-specific prefixes change the KV cache generated for the same shared context, causing each agent to repeatedly prefill the growing context and construct a separate cache with high computation and memory overhead. Selective recomputation reduces this redundancy but still retains substantial model execution, while existing delta correction methods either support only recurring context relations or maintain memory-intensive online correction states for dynamically changing context. For first seen shared context, these methods also construct a reference cache outside the agent workflow, and an approximate correction at the first agent affects the outputs passed to subsequent agents. We present KVCMAS, an online KV cache correction framework that represents cross-agent cache deviations using compact low-rank states and seamlessly chains corrections along the agent workflow without an additional reference prefill. This design supports dynamically changing shared context while preserving an exact first-agent cache. Across multiple language and vision-language workloads, KVCMAS matches or improves the accuracy of prior KV cache sharing methods while achieving the lowest TTFT under highly concurrent serving. Under controlled serving traces, it provides a 2.0x TTFT speedup over inference without KV cache sharing and reduces peak GPU memory by up to 3.7x relative to a prior KV cache correction method. These results establish KVCMAS as an accurate and scalable KV cache sharing approach for prompt-specialized multi-agent serving.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Efficient reasoning and agent inference, one of 12 directions in Efficient LLM inference and serving. Builds on 33 of its 40 references that are in the map. First posted on arXiv 28 Sep 2026.

Builds on

33 of 40 references.

Closest ancestors · 6

Same direction, newest first

  1. 2026Learning Agent Execution for KV-Cache Management in Agentic Serving16 Jul
  2. 2026TraceLab: Characterizing Coding Agent Workloads for LLM Serving29 Jun
  3. 2026TokenDance: Scaling Multi-Agent LLM Serving via Collective KV Cache Sharing3 Apr
  4. 2026RelayCaching: Accelerating LLM Collaboration via Decoding KV Cache Reuse28 Feb
  5. 2025KVCOMM: Online Cross-context KV-cache Communication for Efficient LLM-based Multi-agent Systems14 Oct
  6. 2025KVFlow: Efficient Prefix Caching for Accelerating LLM-Based Multi-Agent Workflows10 Jul

Other references · 27

  1. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  2. 2023MetaGPT: Meta Programming for A Multi-Agent Collaborative Frameworkcited by 1,808
  3. 2023ReConcile: Round-Table Conference Improves Reasoning via Consensus among Diverse LLMscited by 309
  4. 2023Prompt Cache: Modular Attention Reuse for Low-Latency Inferencecited by 268
  5. 2023Unleashing the Emergent Cognitive Synergy in Large Language Models: A Task-Solving Agent through Multi-Persona Self-Collaborationcited by 251
  6. 2023Building Cooperative Embodied Agents Modularly with Large Language Modelscited by 237
  7. 2023Better Zero-Shot Reasoning with Role-Play Promptingcited by 223
  8. 2024CacheBlend: Fast Large Language Model Serving for RAG with Cached Knowledge Fusioncited by 223
  9. 2025MultiAgentBench: Evaluating the Collaboration and Competition of LLM agentscited by 178
  10. 2024Cut the Crap: An Economical Communication Pipeline for LLM-based Multi-Agent Systemscited by 141
  11. 2024Insight-V: Exploring Long-Chain Visual Reasoning with Multimodal Large Language Modelscited by 116
  12. 2024Eigen Attention: Attention in Low-Rank Space for KV Cache Compressioncited by 72
15 more
  1. 2025KVLink: Accelerating Large Language Models via Efficient KV Cache Reusecited by 66
  2. 2024Palu: Compressing KV-Cache with Low-Rank Projectioncited by 64
  3. 2025Cache-to-Cache: Direct Semantic Communication Between Large Language Modelscited by 61
  4. 2025Visual Document Understanding and Reasoning: A Multi-Agent Collaboration Framework with Agent-Wise Adaptive Test-Time Scalingcited by 30
  5. 2025KVShare: An LLM Service System with Efficient and Effective Multi-Tenant KV Cache Reusecited by 22
  6. 2024Triad: A Framework Leveraging a Multi-Role LLM-based Agent to Solve Knowledge Base Question Answeringcited by 18
  7. 2026Cross-Model KV Cache Transfer in LLM Families: A Closed-Form Linear Mapping for Prefill Reusecited by 15
  8. 2025Augmenting Multi-Agent Communication with State Delta Trajectorycited by 14
  9. 2026MetaGen: Self-Evolving Roles and Topologies for Multi-Agent LLM Reasoningcited by 9
  10. 2025MapCoder-Lite: Distilling Multi-Agent Coding into a Single Small LLMcited by 3
  11. 2026GraphFlow: A Graph-Based Workflow Management for Efficient LLM-Agent Servingcited by 3
  12. 2026Characterization of Multi-Model Agentic AI Systems on General Tasks via Trace-Driven Simulationcited by 3
  13. 2026Kamera: Unified Position-Invariant Multimodal KV Cache for Training-Free Reusecited by 3
  14. 2026MA-VLA: Multi-Arm Vision-Language-Action Model for Collaboration and Compositional Generalizationcited by 3
  15. 2026MoRSE: Task-Oriented Multi-Agent System with Mixture of Role-Subtask Expertscited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026LatentRemedy: Efficient Remedy For Non-Prefix KV Cache Reuse in Multi-Agent Systems
  2. 2026PReCache: Efficient KV Cache Sharing for Multi-LoRA Agents via Low-Rank Precomputation and Neutral Reconstruction
  3. 2026LRAgent: Efficient KV Cache Sharing for Multi-LoRA LLM Agents
  4. 2026AgentKVShift: Efficient KV Cache Reuse for Agentic Memory Systems
  5. 2026When KV Cache Reuse Fails in Multi-Agent Systems: Cross-Candidate Interaction is Crucial for LLM Judges
  6. 2026MemDecay: Region-Aware KV Cache Eviction for Efficient LLM Agent Inference
  7. 2025TokenCake: A KV-Cache-centric Serving Framework for LLM-based Multi-Agent Applications
  8. 2026[AAFLOW+] Stateful Operator Abstraction with Zero-Copy Distributed KV Cache Orchestration for Multi-Agent Workflows