LLM Reference

Pangu 3.0 38B

About

The Pangu 3.0 38B model by Huawei is a sophisticated AI language model with 38 billion parameters, designed to enhance industry-specific applications through a "5+N+X" three-layer architecture. Its foundational layer supports domains like NLP, computer vision, and multimodal learning. Trained on 43 years of diverse data, Pangu 3.0 excels in tasks such as natural language understanding and multimodal applications, while its scenario-specific models promote swift deployment. Notably, it powers the Pangu Weather model, achieving unparalleled prediction speed and accuracy. Its versatility spans over 400 business scenarios across more than 10 industries.

Capabilities

MultimodalFunction CallingTool UseJSON Mode

Specifications

FamilyPangu
ArchitectureDecoder Only
Specializationgeneral