NEWSLETTER
Goal has launched the first two models of its multimodal call 4 suite: call 4 scout and call 4 maverick. Maverick is “the battle horse” of the two and stands out in the understanding of the image and text of the “general assistants and the cases of chat use,” said the company in a <a target="_blank" class="no-affiliate-link" href="https://ai.meta.com/blog/llama-4-multimodal-intelligence/”>BlogWhile the smallest Scout model could address things such as “summary of multiple documents, analyze a wide user activity for custom tasks and reasoning on large code bases.” The company also presented Llama 4 Behemoth, an next model that says is “among the smartest LLMs in the world” and CEO <a target="_blank" class="no-affiliate-link" href="https://www.instagram.com/p/DIE0TmPyORV/”>Mark Zuckerberg He said we will listen to a fourth model, call 4 reasoning, “in the next month.”
Both Maverick and Scout are available to download now from the Calls Website and Hugged faceAnd they have added to goal ai, even for WhatsApp, Messenger and instagram DMS.
Scout has 17 billion active parameters with 16 experts, says Meta. According to Zuckerberg, “it is extremely fast, native to multimodal, and has a token context length of 10 million token of 10 million leading, almost infinite, and is designed to function with a single GPU.” Maverick, on the other hand, has 17 billion active parameters with 128 experts. The company says that it exceeds competitors such as GPT-4O and Gemini 2.0 on coding, reasoning, multilinguals, long and contexts and reference images, and compared to Deepseek V3.1 in reasoning and coding.
Zuckerberg is already calling the next Behemoth model, which is still training, “the world's highest performance model”, with 288 billion active parameters, according to the company. It may not yet be here, but we are likely to listen much more about that and the reasoning model soon; The Meta ai developer conference, LlamaCon, is just a few weeks away.
This article originally appeared in Engadget in https://www.engadget.com/ai/meta-introduces-llama-4-with-two-New-models-avilable-now-and- two-mole-on-the -way-214524295.html? SRC = RSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS<script async src="//www.instagram.com/embed.js”>
