\n
\u2192 Scroll horizontally to see all columns<\/p>\n
\n\n\n| Published<\/th>\n | Model<\/th>\n | Where to get it<\/th>\n | Quantizations<\/th>\n | Smallest VRAM tier<\/th>\n | Article<\/th>\n<\/tr>\n<\/thead>\n |
\n\n| 2026-09-26<\/td>\n | Qwen\/Qwen3.8-27B<\/td>\n | lmstudio-community\/Qwen3.8-27B-MLX-4bit<\/a>, lmstudio-community\/Qwen3.8-27B-MLX-8bit<\/a><\/td>\n | MLX 4bit, MLX 8bit<\/td>\n | 24GB<\/td>\n | Qwen3.8-27B Multimodal Vision-Language Model: 8GB+ VRAM, GGUF Builds<\/a><\/td>\n<\/tr>\n\n| 2026-09-18<\/td>\n | prism-ml\/Ternary-Bonsai-2-27B-gguf<\/td>\n | prism-ml\/Ternary-Bonsai-2-27B-mlx-2bit<\/a><\/td>\n | MLX 2bit<\/td>\n | 12GB<\/td>\n | Ternary-Bonsai-2-27B-gguf Text Generation Model: 8GB+ VRAM<\/a><\/td>\n<\/tr>\n\n| 2026-09-17<\/td>\n | harshatheg\/Qwen-2.5-1B-RLCD<\/td>\n | This repository<\/td>\n | \u2014<\/td>\n | 4GB<\/td>\n | Qwen-2.5-1B-RLCD Text Generation Model: 4GB+ VRAM<\/a><\/td>\n<\/tr>\n\n| 2026-09-11<\/td>\n | Edge0\/Edge0-35B-A3B-preview<\/td>\n | This repository<\/td>\n | \u2014<\/td>\n | 24GB<\/td>\n | Edge0-35B-A3B-preview 35B MoE Model for Phone-Class Memory: 24GB+ VRAM<\/a><\/td>\n<\/tr>\n\n| 2026-09-09<\/td>\n | nex-agi\/Nex-N2.5-mini<\/td>\n | mlx-community\/Nex-N2.5-mini-OptiQ-4bit<\/a>, mlx-community\/Nex-N2.5-mini-oQ4<\/a><\/td>\n | MLX 4bit<\/td>\n | 24GB<\/td>\n | Nex-N2.5-mini Agent Model for Long-Horizon Tasks: 16GB+ VRAM<\/a><\/td>\n<\/tr>\n\n| 2026-09-08<\/td>\n | openbmb\/MiniCPM5-2B<\/td>\n | openbmb\/MiniCPM5-2B-MLX<\/a>, mlx-community\/MiniCPM5-2B-8bit<\/a><\/td>\n | MLX, MLX 8bit<\/td>\n | 4GB<\/td>\n | MiniCPM5-2B On-Device Model Strong in Code and Math: 4GB+ VRAM<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n “Smallest VRAM tier” is the smallest tier in the requirements table of each article (for other builds, of those builds; for image, video and audio models, always the article’s own table, which counts every component). Leave headroom for context length.<\/em><\/p>\nLast updated 2026-09-26 (JST). The explanation at the top of this page was written with the help of AI from the primary sources it cites. The tables under “Our Coverage and Data” are assembled by code from our article log.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"What Is MLX? MLX is a machine-learning framework for Apple Silicon (M-series chips), developed by Apple’s machine learning research team. Its companion package mlx-lm runs LLMs, and Hugging Face hosts many models converted and quantized for MLX (for example in the mlx-community organization). When we say “MLX format”, we mean weights prepared to be loaded […]<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-5501","page","type-page","status-publish","hentry"],"lang":"en","translations":{"en":5501,"ja":5500},"_links":{"self":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/pages\/5501","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/comments?post=5501"}],"version-history":[{"count":0,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/pages\/5501\/revisions"}],"wp:attachment":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media?parent=5501"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}} | | | | | |