Text model

Ling 3.0 flash fp4

Ling-3.0-flash is a hybrid-linear Mixture-of-Experts language model with 124 billion total and 5.1 billion activated parameters. It uses a native hybrid linear attention architecture combining Kimi Delta Attention and Gated MLA with sparse MoE routing, enabling efficient long-context processing up to 256K tokens. The model is built for agentic workflows, including coding, tool use, and multi-step reasoning tasks, while activating only a small fraction of its parameters per token.

Type
Text
Weights
Closed
Downloads
15k
Likes
26
Pipeline
text generation
License
mit

Model hub

inclusionAI/Ling-3.0-flash-fp4

Repo
inclusionAI/Ling-3.0-flash-fp4
Access
Public
Technical tags and languages
safetensors bailing_hybrid text-generation conversational custom_code 8-bit fp8
Open on Hugging Face

More from AntGroup

12

About AntGroup

1

Ant Group is a Chinese financial technology company that operates Alipay, one of the world's largest mobile payment platforms with over 1.3 billion users. It provides digital payment, lending, insurance, wealth management, and cloud computing services for consumers and businesses globally.