![\"Meta](\"https:\/\/cdn.iphoneincanada.ca\/wp-content\/uploads\/2025\/04\/meta-ai-hero.jpg\"\/ "\\"meta")
![\"Meta](\"https:\/\/cdn.iphoneincanada.ca\/wp-content\/uploads\/2025\/04\/meta-ai-hero.jpg\" "\\"meta")
<\/p>\n
Meta has released its newest artificial intelligence models under the Llama 4 family<\/a>, and the company says these are its most powerful and advanced models yet.<\/p>\n In a press release on Saturday, Meta introduced Llama 4 Scout and Llama 4 Maverick\u2014two open-source models designed to help developers create smarter AI tools that understand both text and images. Meta also previewed a much larger model still in training, called Llama 4 Behemoth.<\/p>\n Llama 4 Scout is a smaller but highly efficient model that can run on just one high-end graphics card. It\u2019s capable of processing up to 10 million tokens (basically, words or chunks of data) at once, which is a huge jump from previous versions. Meta says it\u2019s faster and more memory-efficient than anything they\u2019ve released before.<\/p>\n As for Llama 4 Maverick, it\u2019s more powerful version aimed at more complex tasks like writing code, solving math problems, and understanding images. Meta says it performs better than some of the top models out there, including OpenAI\u2019s GPT-4<\/a> and Google\u2019s Gemini 2.0<\/a>, all while being more efficient and cost-effective (Gemini 2.5 is already out).<\/p>\n The secret behind these models is a system<\/a> called Mixture-of-Experts, or MoE. Instead of using the entire model for every task, MoE activates only the parts it needs, which makes things run faster and cheaper. Both Scout and Maverick are built using this method.<\/p>\n Meta is also working on an even bigger model, called Llama 4 Behemoth, which has nearly 2 trillion parameters. That model isn\u2019t ready for release yet, but Meta says it\u2019s helping \u201cteach\u201d the other models to be smarter through a method called distillation, which transfers knowledge from one AI to another.<\/p>\n Beyond performance, Meta says it\u2019s focused on making these models safe and balanced. The company has built-in protections to prevent the AI from giving harmful or biased answers, and it has improved how the models respond to politically sensitive or controversial topics. Compared to earlier versions, Llama 4 is less likely to refuse answering a tough question or lean too heavily in one political direction.<\/p>\n The new models are already powering Meta\u2019s AI assistant in WhatsApp<\/a>, Messenger<\/a>, Instagram DMs, and on the Meta.ai website. Developers can also download Llama 4 Scout and Maverick for free through Meta\u2019s website and Hugging Face, a popular platform for AI tools.<\/p>\n Meta says this is just the beginning for the Llama 4 line. The company plans to share more about its future AI plans at LlamaCon, an event happening April 29.<\/p>\n Another day, and another advancement in an AI large language model. It\u2019s hard to keep up with all the progress. Who\u2019s going to be the one to beat in year or two?<\/p>\n <\/div>\n