News

Meta Unveils Llama 3: A 400-Billion Parameter Multimodal Language Model

Meta has unveiled Llama 3, a family of large language models (LLMs) that includes a 400-billion parameter multimodal model, the largest open-source LLM in the world. The company is also integrating its Meta AI chatbot, powered by Llama 3, into the search bar of its four major apps: Facebook, Instagram, Messenger, and WhatsApp.

Llama 3 is a significant step forward for Meta in its competition with OpenAI, which developed the popular ChatGPT LLM. The new model is said to be capable of more complex reasoning and code generation than its predecessors, and it can also process and generate images.

Meta says that Llama 3 was trained on a dataset of over 15 trillion tokens, which is seven times larger than the dataset used to train Llama 2. The company also used four times more code to train Llama 3.

The company has released two smaller versions of Llama 3, an 8-billion parameter model and a 70-billion parameter model, which are both available for free. The 400-billion parameter model is still under development, but Meta plans to release it in the future.

Meta says that one of its main goals for Llama 3 is to reduce the number of times that the AI is unable to answer a question. The company is also working to make Llama 3 more multilingual and to improve its ability to understand and generate different types of content, such as video.

Back to top button