~/AI/edge0-145b-open-source-on-device-ai-with-145b-parameters

Edge0-145B: Open-Source On-Device AI with 145B Parameters

Edge0 has introduced Edge0-145B, the world's first open-source on-device AI model featuring 145 billion parameters. Released on November 12, 2026, alongside Edge0-120B, it brings advanced coding, reasoning, and agent capabilities directly to personal computers. This release challenges the paradigm that massive frontier-class models require centralized cloud datacenters to operate. By enabling high-parameter models locally on standard consumer hardware, it enhances user privacy, reduces cloud dependency, and pushes the limits of edge computing. The model utilizes innovative architecture features such as Latent Mixture of Experts (MoE), SSD-streamed weights, and a predictive load router to run efficiently on a single GPU or laptop. Developers can access the weights on Hugging Face alongside specialized inference tools like edge0-infer and edge0-strata on GitHub.

## BACKGROUND

On-device AI refers to running artificial intelligence models locally on consumer devices like laptops or smartphones rather than remote servers. Large Language Models (LLMs) with high parameter counts traditionally required heavy server-grade hardware, but compression and streaming techniques now make local execution feasible.

## REFERENCES

## KEYWORDS

#AI#Large Language Models#Open Source#Edge Computing

$ subscribe --daily