株式会社オブライト
AI Model2026-05-17

Llama (Meta)

Also known as: Llama (Meta) / Llama 4 / Meta Llama

Meta's open-weights large language model family. The Llama 4 generation adopts MoE architecture and ships under a commercial-friendly licence enabling local deployment and fine-tuning.


Overview

Llama 4 uses a Mixture-of-Experts design for efficient inference and is a leading local-LLM option alongside Gemma 4 and Qwen 3.5, easily run via Ollama. See Gemma 4 vs Llama 4 vs Qwen 3.5.

Ecosystem

Supported by Meta AI, Ray, vLLM, and many other frameworks, making it one of the most widely deployed open models in on-premises enterprise AI.

Related Columns

Related Terms

Feel free to contact us

Contact Us