GLM 5.2 open model rivals frontier closed models

The open-source model runs at 120 tok/s on two Blackwell boxes and tops Terminal Bench 2.1. Users compare it to Opus 4.8 and GPT-5.5. It's available on HuggingFace and supported by Perplexity, HuggingChat, and NVIDIA's NVFP4 quantization.