Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
Meta has launched its new AI model, Llama 3, which includes 8B and 70B parameter versions, enhancing its capabilities in multiple languages and various AI applications. The model, available on major platforms like AWS, Google Cloud, Microsoft Azure, Snowflake, IBM WatsonX, Nvidia NIM, and integrated with Hugging Face and Meta AI assistant, supports over 30 languages and has been integrated into Meta's apps such as Facebook, Instagram, and WhatsApp. Llama 3 is part of Meta's broader strategy to lead in the generative AI space, competing directly with models like OpenAI's ChatGPT. Additionally, Meta is already developing a more advanced model, Llama 4, with 400 billion parameters, promising even greater performance.
"we have gotten more optimistic and ambitious in AI" says Zuckerberg. Llama 3/latest models are "leading models" and Zuck says Meta can be the "leading AI company in the world" and that will open up new opportunities. "We have the talent, the data and the ability to scale the…
A 65,000 token context window on open source LLaMA 3 is an absolute game changer for local AI. It will change just about everything. I know 100s of corporate clients that would jump to it in a heartbeat. Astonishing work by Wing and team! https://t.co/0gqncamSRW
Meta’s commitment to sharing its AI software from the beginning has been a big draw for talented AI researchers, who have in turn improved the quality of its models. https://t.co/einnRZToAh AI Agenda by @steph_palazzolo
