{"slug": "show-hn-kv-cache-grafting-boosting-frozen-12b-llms-to-93-3-aime-accuracy", "title": "Show HN: KV-Cache Grafting – Boosting frozen 12B LLMs to 93.3% AIME accuracy", "summary": "Researchers introduced KV-Cache Grafting, a technique that deposits verified knowledge as byte-exact key-value state artifacts and grafts them into inference contexts, boosting a frozen 12B Gemma-4 model from 80.0% to 93.3% on AIME 2025 without weight changes. The method achieves bit-exact logits, zero KL divergence, and reduces token usage by up to 6,574x, while extending usable context from 32,768 to 2,854,766 tokens at zero extra memory.", "body_md": "# Computer Science > Computation and Language\n\n[Submitted on 15 Jul 2026]\n\n# Title:Smarter and Cheaper at Once: Byte-Exact KV-Cache Grafting Turns a Frozen Small Model into a Verified-Knowledge Flywheel\n\n[View PDF](/pdf/2607.14431)\n\n[HTML (experimental)](https://arxiv.org/html/2607.14431v1)\n\nAbstract:We report a way to make a frozen small language model both more capable and dramatically cheaper at once, without changing any weights. Verified knowledge is deposited once as a byte-exact key-value (KV) state artifact and later restored, by graft, into a fresh inference context. The restore is bit-exact: under a pinned deterministic configuration, the grafted logits are byte-for-byte identical to a fresh computation (SHA-256 equality), with zero KL divergence and 100% argmax agreement over fifty samples. We show that own-position graft is the unique numerically exact operating point on a model with floating-point rotary encoding, and we verify byte-exactness on two model scales (12B, 31B) and two GPU targets, one through a pre-registered replay. On AIME 2025, a frozen Gemma-4-12B moves from 80.0% to 93.3% once a verified solution library is grafted, above its own 77.5% and its 31B sibling's 89.2% published anchors. On the recurring case, eight problems the base model never solves within a 401,026-token budget are answered from cached verified solutions in 61 total decode tokens, a factor of 6,574 fewer tokens and about 8,700x less energy; the capability claim proper rests on held-out transfer (7 of 7 at 31B). The same byte-exact store widens usable context from 32,768 to 2,854,766 tokens at zero extra accelerator memory, and moves byte-identical between machines of the same architecture. We describe the system at the behavior level; the engine is proprietary, and every reported number is backed by committed input and output hashes so the scoring can be re-checked without it.\n\n### Current browse context:\n\ncs.CL\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/))# 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))# 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))# 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/show-hn-kv-cache-grafting-boosting-frozen-12b-llms-to-93-3-aime-accuracy", "canonical_source": "https://arxiv.org/abs/2607.14431", "published_at": "2026-07-17 02:37:51+00:00", "updated_at": "2026-07-17 02:51:10.405808+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-research", "ai-infrastructure"], "entities": ["Gemma-4", "AIME 2025"], "alternates": {"html": "https://wpnews.pro/news/show-hn-kv-cache-grafting-boosting-frozen-12b-llms-to-93-3-aime-accuracy", "markdown": "https://wpnews.pro/news/show-hn-kv-cache-grafting-boosting-frozen-12b-llms-to-93-3-aime-accuracy.md", "text": "https://wpnews.pro/news/show-hn-kv-cache-grafting-boosting-frozen-12b-llms-to-93-3-aime-accuracy.txt", "jsonld": "https://wpnews.pro/news/show-hn-kv-cache-grafting-boosting-frozen-12b-llms-to-93-3-aime-accuracy.jsonld"}}