{"slug": "cache-to-cache-direct-semantic-communication-between-large-language-models", "title": "Cache-to-Cache: Direct Semantic Communication Between Large Language Models", "summary": "Researchers Tianyu Fu and colleagues posted Cache-to-Cache (C2C) to arXiv on 3 Oct 2025, revised 2 Mar 2026, a method that lets large language models communicate directly through fused KV-cache representations instead of text. C2C uses a neural network to project and fuse the source model's KV-cache into the target model's, with a learnable gating mechanism selecting which target layers receive the transfer. The authors report C2C achieves 6.4-14.2% higher average accuracy than individual models, beats text-based communication by roughly 3.1-5.4%, and delivers an average 2.5x latency speedup, with code at github.com/thu-nics/C2C.", "body_md": "# Computer Science > Computation and Language\n\n  [Submitted on 3 Oct 2025 (\n\n[v1](https://arxiv.org/abs/2510.03215v1)), last revised 2 Mar 2026 (this version, v2)]\n# Title:Cache-to-Cache: Direct Semantic Communication Between Large Language Models\n\n[View PDF](https://arxiv.org/pdf/2510.03215)\n\n[HTML (experimental)](https://arxiv.org/html/2510.03215v2)\n\nAbstract:Multi-LLM systems harness the complementary strengths of diverse Large Language Models, achieving performance and efficiency gains that are not attainable by a single model. In existing designs, LLMs communicate through text, forcing internal representations to be transformed into output token sequences. This process both loses rich semantic information and incurs token-by-token generation latency. Motivated by these limitations, we ask: Can LLMs communicate beyond text? Oracle experiments show that enriching the KV-Cache semantics can improve response quality without increasing cache size, supporting KV-Cache as an effective medium for inter-model communication. Thus, we propose Cache-to-Cache (C2C), a new paradigm for direct semantic communication between LLMs. C2C uses a neural network to project and fuse the source model's KV-cache with that of the target model to enable direct semantic transfer. A learnable gating mechanism selects the target layers that benefit from cache communication. Compared with text communication, C2C utilizes the deep, specialized semantics from both models, while avoiding explicit intermediate text generation. Experiments show that C2C achieves 6.4-14.2% higher average accuracy than individual models. It further outperforms the text communication paradigm by approximately 3.1-5.4%, while delivering an average 2.5x speedup in latency. Our code is available at [this https URL](https://github.com/thu-nics/C2C).\n    \n\n## Submission history\n\nFrom: Tianyu Fu [\n[view email](https://arxiv.org/show-email/5bc6748f/2510.03215)]\n\n**Fri, 3 Oct 2025 17:52:32 UTC (484 KB)**\n\n[\\[v1\\]](https://arxiv.org/abs/2510.03215v1)\n**[v2]** Mon, 2 Mar 2026 19:24:02 UTC (546 KB)\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer \n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers \n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps \n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations \n\n*(*[What are Smart Citations?](https://www.scite.ai/))\n# Code, Data and Media Associated with this Article\n\nalphaXiv \n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers \n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub \n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub \n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face \n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast \n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))\n# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower \n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender \n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))\n# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [**Learn more about arXivLabs**](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/cache-to-cache-direct-semantic-communication-between-large-language-models", "canonical_source": "https://arxiv.org/abs/2510.03215", "published_at": "2026-09-18 18:55:35+00:00", "updated_at": "2026-09-18 19:24:20.219191+00:00", "lang": "en", "topics": ["large-language-models", "ai-research", "machine-learning", "natural-language-processing"], "entities": ["Cache-to-Cache", "Tianyu Fu", "arXiv", "KV-Cache", "thu-nics"], "alternates": {"html": "https://wpnews.pro/news/cache-to-cache-direct-semantic-communication-between-large-language-models", "markdown": "https://wpnews.pro/news/cache-to-cache-direct-semantic-communication-between-large-language-models.md", "text": "https://wpnews.pro/news/cache-to-cache-direct-semantic-communication-between-large-language-models.txt", "jsonld": "https://wpnews.pro/news/cache-to-cache-direct-semantic-communication-between-large-language-models.jsonld"}}