Photo credit: www.cnbc.com
On January 29, 2025, visitors were seen capturing images near the sign in front of Meta’s headquarters located in Menlo Park, California.
Meta has recently rolled out the initial models of its new open-source AI software, Llama 4, as the tech giant accelerates its efforts to dominate the generative AI landscape.
However, the company has not yet introduced its most advanced iteration, the Llama 4 Behemoth model, which is anticipated to surpass its competitors in functionality and serve as the foundational model for the new AI offerings. This powerful version is reportedly still in the training phase, according to details shared in Meta’s official blog post.
Meta’s Llama 4 models are designed to enhance AI agents that exhibit advanced reasoning capabilities and can perform comprehensive tasks. As described by Chris Cox, Meta’s chief product officer, these agents will have the ability to navigate the internet and assist both consumers and businesses with various needs.
Two of the newly unveiled models, Llama 4 Scout and Llama 4 Maverick, are now accessible for users via Meta AI within platforms like WhatsApp, Messenger, Instagram Direct, and the Meta AI website.
In a recent video shared on Instagram, Meta CEO Mark Zuckerberg expressed, “Our ambition is to create the world’s leading AI, make it open source, and ensure it is available for the benefit of everyone globally.” He added, “I have long believed that open-source AI will emerge as the dominant paradigm, and with Llama 4, we are beginning to see that come to fruition. Meta AI is undergoing a significant upgrade today.”
Looking ahead, Meta is set to launch its inaugural LlamaCon AI conference on April 29. Additionally, reports anticipate the introduction of a standalone app dedicated to the Meta AI chatbot in the second quarter of the year, as indicated by CNBC in February.
Source
www.cnbc.com