Meta Llama 3

Meta released Llama 3 on my birthday! 🎂 Best present ever, thanks Meta! 😀

— Andrew Ng (@AndrewYNg) April 18, 2024

Nice benchmark results for Llama-3 8B and 70B. https://t.co/SJEbwOOE9w

— Yann LeCun (@ylecun) April 18, 2024

• Meta Llama 3 Release Meta has released the first two models of Llama 3, featuring 8B and 70B parameters, with state-of-the-art performance and new capabilities like improved reasoning.

• Open Source and Community Focus Llama 3 is open source, aiming to spur innovation in AI applications, developer tools, evaluations, and inference optimizations.

• Enhanced Model Performance The new models surpass Llama 2 with advancements in pretraining, post-training, and instruction fine-tuning, resulting in reduced false refusals and more diverse responses.

• Future Developments Plans for Llama 3 include multilingual and multimodal capabilities, longer context, and continued performance improvements across core LLM capabilities.

Introducing Meta Llama 3: The most capable openly available LLM to date