Meta introduced Artificial Intelligence Model Llama 4!

Meta opened the door of a new era in the field of artificial intelligence with Llama 4. The LLAM 4 model, which is introduced at the launch of the company at the weekend, consists of three different structures: Llama 4 Scout, Llama 4 Maverick and Study 4 Behemoth.

The Llama 4 family attracts attention not only on the text but also with visual and video data. Scout and Maverick models were opened to access via meta’s official platform llalama.com. The Behemoth is not yet open to everyone. It is stated that this model is used only in internal tests.

All of the new models are integrated with Meta’s own artificial intelligence -supported applications. WhatsApp, Messenger and the Meta AI assistant on Instagram have been updated to use Llama 4 in 40 countries. Multimodal features will now be active only in the US and English language.

Meta uses the architecture of “Mixure of Experts for the first time in Llama 4 family. This architecture works by dividing its large model structure into a large number of tasks -oriented sub -model. Thus, according to the request from the user, only relevant sub -models are run, which increases the efficiency of the system.

For example, Llama 4 Maverick, which has a total parameter of 400 billion, actively uses only 17 billion parameters. Meta Management, with new models GPT-4O and Gemini 2.0, such as competitors in front of their competitors said.

So what do you think about this? You can share your opinions with us in the comments section below.