MiniMax M3.1 Flash 与 Unsloth 合作，428B 模型量化至 143.9GB
MiniMax 的 428B 大模型被 Unsloth 压到 143.9GB 还能保持 91.5% 的输出一致率，想本地跑开源模型的朋友可以试试。
MiniMax 宣布与 Unsloth 社区合作，为 M3.1 Flash 提供 Dynamic GGUF 量化版本。这个 428B 参数的 MoE 模型被压缩到 143.9GB 的 2-bit 体积，Unsloth 评估显示在七个领域中保持 91.5% 的 top-1 token 一致率。用户可通过 Unsloth Desktop 和 llama.cpp 自部署，还能用 OpenAI 兼容 API 接入现有应用。Unsloth 在 GitHub 有 78K+ stars，Hugging Face 下载量超过 5 亿次。
MiniMax M3.1 Flash meets the @UnslothAI community. 🐙🦥✨
Unsloth’s Dynamic GGUFs bring our 428B MoE model to a 143.9GB 2-bit footprint, with 91.5% top-1 token agreement in their evaluation across seven domains.
That means a lower memory barrier to self-hosting, deployment through Unsloth Desktop and llama.cpp, and an OpenAI-compatible API for existing apps and agents.
With 78K+ GitHub stars and over 500M model downloads on Hugging Face, Unsloth gives M3.1 Flash a path into an established community of developers running, evaluating and building with open models.
Open weights. Local control. More ways to build.
