Meta Announces New Llama 4 AI Models to Compete with ChatGPT 4 and Google’s Gemini
Meta Platforms Launches Llama 4 AI Models
Meta Platforms, Inc. (stock ticker: META) has recently introduced the initial models of its cutting-edge open-source artificial intelligence (AI) software, known as Llama 4. This significant announcement was made over the weekend.
Overview of the Llama 4 Models
Meta’s latest offering consists of two distinct models: Llama 4 Scout and Llama 4 Maverick. Both of these models boast 17 billion active parameters, but they differ in their architecture. The Scout model features 16 experts, while the Maverick model includes 128 experts. Notably, these new Llama 4 models are the first to utilize a mixture of experts (MoE) architecture, which aims to improve AI performance and versatility.
Despite the excitement surrounding the release, the most powerful variant, Llama 4 Behemoth, has yet to be launched as it is still undergoing training, according to the company’s blog update.
Enhanced Capabilities and Accessibility
Chris Cox, Chief Product Officer at Meta, emphasized that Llama 4 is designed to elevate the capabilities of AI agents, enhancing their reasoning and action skills. These advancements can be beneficial for both consumers and businesses, as the AI agents will be able to handle various tasks, including web surfing and other applications.
Mark Zuckerberg highlighted the company’s mission in a recent Instagram video, stating, "Our goal is to build the world’s leading AI, open source it, and make it universally accessible so that everyone in the world benefits." This reveals Meta’s commitment to making AI innovations available to a broader audience.
Users can explore the new Llama 4 models through various platforms, including WhatsApp, Messenger, Instagram Direct, and the Meta AI website.
Upcoming Events and Announcements
Meta is also set to host the first LlamaCon AI conference on April 29. During this event, attendees can expect further announcements, including the planned introduction of a dedicated app for the Meta AI chatbot, anticipated in the second quarter of the year.
Why Llama 4 Matters
Meta has lauded the Llama 4 releases as its "most advanced models yet," asserting their superiority in multimodality. The company claims that these models outperform existing competitors such as ChatGPT 4, Google’s Gemini 3 and Gemini 2.0 Flash-Lite, and Mistral 3.1 in various aspects, suggesting that Llama 4 could be a major player in the AI landscape.
However, the launch comes at a critical juncture for Meta, as the company is experiencing significant shifts in its AI research leadership. Joelle Pineau, Vice President of AI Research at Meta, recently announced her departure from the company, effective in May.
Controversies and Challenges
Meta’s AI models have been the subject of scrutiny, particularly in light of recent legal challenges. A court filing in the Kadrey v. Meta lawsuit revealed that Meta had allegedly profited from its Llama AI models through revenue-sharing agreements with third-party businesses, contradicting earlier claims that downplayed such practices.
Despite these challenges, Meta remains steadfast in its investment in AI technology. The recent launch of Llama 4 models marks the latest endeavor in this ongoing effort.
Current Stock Performance
According to Benzinga’s Proprietary Edge Rankings, Meta holds a momentum rating of 57.81% and a growth rating of 74.76%. These ratings are based on the stock’s historical earnings and revenue growth over various timeframes, indicating a focus on both long-term trends and recent performance.
As of Friday, shares of Meta closed at $504.73, reflecting a decline of 5.06% and a drop of over 19% in the past month.
