Meta Launches Llama 3 AI Models in 8B and 70B, Enhances App Integration

Apr 18, 2024, 08:30 PM

Meta Platforms has released its latest large language model, Llama 3, which includes a range of models with varying capabilities. The new models, available in 8B and 70B parameter sizes, have been trained on 15 trillion tokens and boast a context length of 8,000 tokens. This release marks a significant improvement over Llama 2, with the new models trained on a dataset seven times larger than its predecessor. Llama 3 models are designed to enhance language understanding and task performance across a wide range of applications, and they are already integrated into Meta's various apps including Facebook, Instagram, WhatsApp, and Messenger. Additionally, a 400B+ model is still in training and promises to deliver even more advanced capabilities. The Llama 3 models are also available on Azuremarketplace and have been integrated into the search functions across Meta's apps, including the new Meta AI chatbot.