On Saturday, Meta Platforms released the latest version of its language model known as Llama 4, specifically Llama 4 Scout and Llama 4 Maverick.

According to Reuters, Meta claims that Llama is a multimodal AI system.

These systems can process and integrate various data types, including text, video, images, and audio, and transform content across these formats.

Meta also stated that Llama 4 Maverick and Llama 4 Scout will feature open-source software.

Additionally, the company announced a preview of Llama 4 Behemoth, referring to it as “one of the smartest LLMs in the world and our most powerful, serving as a teacher for new models.”

Major tech firms are heavily investing in AI infrastructure following the success of OpenAI's ChatGPT, which has reshaped the technology landscape and spurred investments in machine learning.

On Friday, The Information reported that Meta delayed the launch of the latest LLM version, as Llama 4 did not meet the company’s expectations during technical testing, particularly in reasoning and mathematical tasks.

Concerns were also raised that Llama 4 was less capable than OpenAI models in conducting human-like voice conversations, the report added.