@w_o_w 在 mind lab开源了一个基于GLM-5.1后训练的模型Macaron-V1-Preview，多项Benchmark超过gpt5.4 中发帖
[image]
[image]
mindlab-research/Macaron-V1-Preview-749B · Hugging Face — mindlab-research/Macaron-V1-Preview-749B · Hugging Face