NVIDIA launches GLM-5.2 open-weight model

Hugging Face Blog

GLM-5.2 is a sub-trillion parameter open-weight model from NVIDIA, available for free via Hugging Face Inference Providers. Early users report it achieves 80% on a financial benchmark and runs at 120 tok/s on two Blackwell tinyboxes, rivaling frontier models like Opus 4.8.

glm-5-2NvidiaOpen source

See 14 more sources

zai-org/GLM-5.212 days agozai-org

nvidia/GLM-5.2-NVFP42 days agonvidia

unsloth/GLM-5.2-GGUF11 days agounsloth

zai-org/GLM-5.2-FP811 days agozai-org

GLM is the kind of model that revives serious interest in open source AI. It passes the blind test...6 days agoAravind Srinivas

RT @MaziyarPanahi: Got GLM-5.2 running on my Mac Studio via llama.cpp, the reasoning behind all my...1 day agoHugging Face

RT @victormustar: Open source MUST win 🔥 GLM-5.2 is free when used with Hugging Face Inference...10 days agoHugging Face

RT @NielsRogge: Impressive release by @ornith_! Getting the top spot on Terminal Bench 2.1 with...2 days agoHugging Face

China's Zhipu is closing in on top U.S. AI models with Anthropic and OpenAI held back2 days ago

GLM-5.2 and IndexShare for Long-Context Sparse Attention10 days agoSebastian Raschka

[AINews] not much happened today8 days ago

[AINews] GLM-5.2: the top Frontend Coding model in the world, IndexShare for Speculative Decoding11 days ago

GLM-5.2 Is The New Best Open Model6 days agoZvi Mowshowitz

Liminality5 days ago

Hugging Face Blog

Jun 22, 7:55 PM