精选理由
想省显存又不想降精度？GLM-5.2的NVFP4版在vLLM上线了，比FP8省一半内存，推理编码长文本都稳。
NVIDIA发布GLM-5.2的NVFP4检查点，在Blackwell GPU上相比FP8内存占用降低一半。该模型在推理、编码和长上下文基准测试中保持与FP8相同的准确率。用户可通过vLLM直接加载运行：vllm serve nvidia/GLM-5.2-NVFP4。
原文 · vLLM
GLM-5.2 in NVFP4 is ready to serve in vLLM 🚀 @NVI…
GLM-5.2 in NVFP4 is ready to serve in vLLM 🚀
@NVIDIAAI's official NVFP4 checkpoint of GLM-5.2 on Blackwell cuts the memory footprint vs FP8 while matching its accuracy across reasoning, coding, and long-context benchmarks.
Serve it today with: vllm serve nvidia/GLM-5.2-NVFP4
🤗 https://t.co/utHk49BULz