GLM-5.2-FP8: FP8 quantized model published on Hugging Face
The FP8 quantized version of GLM-5.2 was published on Hugging Face under the zai-org organisation. The model is a text-generation transformer with MoE architecture, supporting English and Chinese, licensed MIT, and deployable on Azure and SageMaker.
PUBLISHED2026-06-16
OBSERVED2026-08-11
AGE2mo
SOURCES1
- Published on Hugging Face under
zai-org/GLM-5.2-FP8 - Pipeline: text-generation
- Library: transformers
- Architecture: MoE (glm_moe_dsa)
- Precision: FP8
- Languages: English, Chinese
- License: MIT
- References arxiv:2602.15763, arxiv:2603.12201
- Tags: endpoints_compatible, deploy:azure, deploy:sagemaker, region:us
COMMUNITY
No curated reactions recorded for this event. Facts and takes are kept in separate layers — community context is added by hand, never blended into the record above.