Meta Unveils Llama 4 AI Models, Expands Open Source Offerings
- tech360.tv
- 17 hours ago
- 1 min read
Meta Platforms has launched the latest versions of its large language model, Llama, introducing Llama 4 Scout and Llama 4 Maverick as its most advanced artificial intelligence systems to date.
The company described the new models as multimodal, capable of processing and integrating text, video, images and audio, and converting content across these formats.
Meta said both Llama 4 Scout and Llama 4 Maverick will be released as open source software.
The company also previewed Llama 4 Behemoth, which it called one of the smartest large language models in the world and its most powerful yet. Meta said Behemoth will serve as a teacher for future models.
The release comes amid intense competition in the AI sector, with major technology firms investing heavily in infrastructure following the rise of OpenAI’s ChatGPT.
Meta plans to spend up to USD 65 billion this year to expand its AI infrastructure, as investors push for returns on AI investments.
According to a report by The Information, Meta delayed the launch of Llama 4 after the model failed to meet internal benchmarks, particularly in reasoning and mathematics.
The report also noted concerns that Llama 4 was less capable than OpenAI’s models in conducting humanlike voice conversations.
Meta released Llama 4 Scout and Llama 4 Maverick as open source AI models
The models are multimodal, handling text, video, images and audio
Llama 4 Behemoth previewed as Meta’s most powerful model
Source: REUTERS