{"version":1,"type":"story","url":"https://digestai.news/story/the-kaitchup-releases-statistically-lossless-qwen3-8-27b-gguf-models","json":"https://digestai.news/story/the-kaitchup-releases-statistically-lossless-qwen3-8-27b-gguf-models.json","markdown":"https://digestai.news/story/the-kaitchup-releases-statistically-lossless-qwen3-8-27b-gguf-models.md","slug":"the-kaitchup-releases-statistically-lossless-qwen3-8-27b-gguf-models","headline":"The Kaitchup releases statistically-lossless Qwen3.8-27B GGUF models","summary":"The Kaitchup published three quantized versions of the Qwen3.8‑27B large language model using a customized Statistically‑Lossless Quantization (SLQ) method and packaged them as GGUF files. The three releases target Expected Acceptance Rates (EAR) of 98 %, 98.5 % and 99 % measured on the BF16 reference’s top‑10 tokens, and differ in precision choices such as Q4\\K/Q8\\0 for EAR 98 and Q6\\K/BF16 for EAR 99. Each file keeps 351,008,768 bytes of quantized MTP tensors from the UD‑Q5\\K\\XL donor, excludes vision tensors, and is licensed under Apache 2.0.\n\nValidation was performed with a llama.cpp build on an NVIDIA GH200 GPU (commit 45451c366d4256f94fbc8845645e31a6b9155054), confirming native text scoring and GGUF parity, though image/video inference and speculative decoding were not covered. The releases have not been benchmarked for downstream task performance, and statistical equivalence remains unproven. The models can be downloaded via the Hugging Face CLI, and a simple OpenAI‑compatible server command is provided. The files saw 239 downloads in the last month, and the author invites users to support the work.","keyPoints":["The Kaitchup released three Qwen3.8‑27B GGUF models quantized with SLQ, targeting EAR 98 %, 98.5 % and 99 %.","Models use Q4/K, Q5/K, Q6/K precision combos and retain 351,008,768 bytes of quantized MTP tensors; no vision tensors included.","Validation on an NVIDIA GH200 with llama.cpp confirmed text parity; downstream task equivalence not yet established; 239 downloads last month."],"whyItMatters":"High‑fidelity, low‑size quantized models let developers run a 27B LLM on modest hardware, expanding open‑source AI accessibility.","category":{"slug":"models","name":"Generative AI & Models","url":"https://digestai.news/category/models"},"entities":{"companies":["The Kaitchup","Unsloth","ISTA-DASLab","Hugging Face"],"models":["Qwen3.8-27B","Qwen3.8-27B-SLQ-EAR98-Q5KS","Qwen3.8-27B-SLQ-EAR98.5","Qwen3.8-27B-SLQ-EAR99"],"people":["Michael Helcig","Eldar Kurtic","Dan Alistarh"]},"firstPublishedAt":"2026-10-27T18:52:27Z","updatedAt":"2026-10-27T18:52:27Z","sourceCount":1,"hasPrimarySource":true,"sources":[{"outlet":"huggingface.co","title":"Statistically Lossless Qwen3.8 27B GGUFs","url":"https://huggingface.co/kaitchup/Qwen3.8-27B-SLQ-GGUF","publishedAt":"2026-10-27T18:52:27Z","type":"primary","primary":true,"lead":true}],"sourceNotes":null,"discussions":[{"site":"Reddit","url":"https://www.reddit.com/r/LocalLLaMA/comments/1xh17tk/statistically_lossless_qwen38_27b_ggufs/","points":null}],"thread":{"title":"AtomicChat's Quantized Model Rollout Saga","url":"https://digestai.news/thread/atomicchat-releases-quantized-ling-3-1-flash-models-with-8688-token-accuracy","storyCount":3},"cite":{"text":"Digest AI, \"The Kaitchup releases statistically-lossless Qwen3.8-27B GGUF models\", 27 October 2026, https://digestai.news/story/the-kaitchup-releases-statistically-lossless-qwen3-8-27b-gguf-models","publisher":"Digest AI","title":"The Kaitchup releases statistically-lossless Qwen3.8-27B GGUF models","datePublished":"2026-10-27T18:52:27Z","url":"https://digestai.news/story/the-kaitchup-releases-statistically-lossless-qwen3-8-27b-gguf-models"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}