腾讯混元发布Hy3 1-bit/4-bit版本，295B模型单GPU可运行
Can't Find Benchmark and download link ? See here 👇https://t.co/2bYLI8HRus
腾讯把295B的模型压到1-bit和4-bit，单卡就能跑，用llama.cpp就能玩，赶快试试。
腾讯混元发布了Hy3的1-bit和4-bit量化版本，该模型拥有295B参数。通过1-bit和4-bit量化，模型可在单张GPU上运行，大幅降低硬件门槛。
支持使用llama.cpp进行推理，并启用MTP（多token预测）功能。
模型以GGUF格式提供下载，用户可快速体验。
Can't Find Benchmark and download link ? See here 👇https://t.co/2bYLI8HRus
Can't Find Benchmark and download link ? See here � huggingface.co/AngelSlim/Hy3-… s Tencent Hy @TencentHunyuan We’ve just released the 1-bit & 4-bit version of Hy3, a flagship-scale 295B model that can be served on a single GPU. 👌 Run Hy3 with llama.cpp, enable MTP, and experience powerful intelligence on dramatically lower hardware.🚀🚀🚀 Can’t wait to see what you buil #Hy3 # #Hy #GGUF G #llamacpp acpp 🔗 View Quoted Tweet📊 3 ⚡
This report is categorized under AITOP's "Models" desk, focusing on China's AI industry and domestic foundation model ecosystem. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Hy3, TencentHunyuan, quantized models, llama.cpp.
Source attribution: curated from Hunyuan (@TXhunyuan).
