Meta Unveils Llama 3 as the Most Capable Open Source Language Model
Meta releases Llama 3, an advanced open source large language model available in 8B and 70B parameter sizes. The new model achieves state-of-the-art performance and powers the upgraded Meta AI assistant.
Meta officially introduces Llama 3, the next generation of its open source large language model, as the most capable openly available LLM to date. The initial release includes pretrained and instruction-fine-tuned models with 8B and 70B parameters that support a wide range of use cases. Llama 3 demonstrates state-of-the-art performance on various industry benchmarks and delivers improved reasoning capabilities.
The new model powers Meta AI, which now stands as one of the world's leading AI assistants for learning, content creation, and task management. Broad availability for Llama 3 spans across major cloud platforms including AWS, Google Cloud, Microsoft Azure, and Hugging Face. Additionally, hardware support comes from prominent chipmakers such as AMD, Intel, NVIDIA, and Qualcomm.
Meta develops Llama 3 with a strong focus on responsibility, introducing new trust and safety tools like Llama Guard 2, Code Shield, and CyberSec Eval 2. The company plans to share the Llama 3 research paper and release additional capabilities in the coming months, including longer context windows, more model sizes, and further performance enhancements.