Mark Zuckerberg’s Meta on Saturday launched its latest generation of Llama (large language model) artificial intelligence models- Scout, Maverick and Behemoth — as the competition in the AI space heats up between Google, OpenAI and other contenders.
In a blog post, the company revealed that the three models would enable users to build more “personalised multimodal experiences” and claimed that they outperformed Gemma 3, Gemini 2.0, Flash Lite and Mistral 3.1 across a wide range of benchmarks.
Meta also claimed that the Llama 4 Scout was the best multimodal AI model currently available in the world. All three of the AI models launched are equipped to tackle almost all tasks, ranging from document summarisation to advanced reasoning across text, images, and video.
The announcement comes after the race to dominate the open-source AI world has heated up with new entrants like China’s DeepSeek toppling heavyweights like OpenAI.
After DeepSeek’s R1 and V3 models outperformed Llama-2, Meta reportedly accelerated the development of their Llama-4 suite and introduced “war rooms” to reverse engineer the source of DeepSeek’s high efficiency, low cost models, reported Firstpost.
The company stated that Llama-4’s Behemoth was “one of the smartest LLMs in the world and our most powerful yet to serve as a teacher for our new models.”
Meta also claims that Maverick, its flagship AI assistant, can outperform OpenAI’s GPT-4o and Google’s Gemini 2.0 on a range of coding, reasoning, and image-based benchmarks.
Scout and Maverick are freely available on Meta’s website, while Behemoth is still in a preview stage.
