Meta Ai Lalama 4: Mark Zuckerberg-Led Meta has Announced Lalama 4, its newest collection of ai models that now power the meta ai assistant on the web and in whatsapp, messenger, messeenger, and inseagram. The AI Models to Compete Against Google’s Gemini and Openai’s chatgpt.
The company introduced the first two models from its multimodal llama 4 suite: llama 4 scout and lLAma 4 Maverick. Both the lLAma 4 scout and lLAma 4 Maverics models are available for download on the meta website and hugging face.
The new models draw inspiration from chinese ai startup deepsek, adopting a machine learning approach approach as mixture of experts. This Technique Lets Different Parts of the Model Specialize in Specific Tasks, Leading to Improved Performance and Greater efficiency.
Today is the start of a new era of naturally multimodal ai innovation.
Today, We’re introducing the first llam 4 models: lLAma 4 scout and llam 4 Maverick – Our Most Advanced Models Yet and the Best in his class for multimodality.
LLAma 4 Scout • 17b-Active-Parameter Model… pic.twitter.com/Z8P3H0MA1P-AI at Meta (@Aaita) April 5, 2025
The lLAma 4 scout is the lighter, more compact model, powerful enough to run on a single nvidia h100 gpu. It is designed for efficiency without sacrificing performance. Ingtrast, LLAMA 4 Maverick is the Heavy Hitter-More Comparable to Advanced Models Like GPT-4o and Gemini 2.0 Flash.
Meta has trained its llama 4 models to be naturally multimodal by pre-training them on vast amounts of unlabeled text, image, and video data. In simple terms, this means the models can undersrstand and respond to both images and text seamlessly. Hence, they are billt to handle conversations that involve more than just words.
On the other hand, the meta is still training llam 4 behemoth, which ceo mark zuckerberg proudly calls “The Highest performing base model in the world.” Internally, meta referrs to it as one of the smartest large language models (LLMS) Ever Created.
According to a company, The LLAma 4 Maverick Model is “The Worster” of the two and excels at image and text understanding for “General Assistant and Chat Use Cases”.
LLAma 4 Scout Comes With 17 Billion Active Parameters, Supported By 16 Experts, Bringing The Total Parameter Count to 109 Billion. LLAMA 4 MAVERICK ALESO FEATURES 17 Billion Active Parameters but Taps INTO A MUCH LARGER POOL OF 128 Experts, Giving it a boost in performance and versatitiy.
However, the multimodal features of meta ai are currently limited to English users in the US, meaning no ghibli style image generation with meta ai just anniversary.