Meta has launched Llama 3.1 405 billion, a large language model with 405 billion parameters, in public preview.

Llama 3.1 405 billion is reportedly capable of tasks like general knowledge, math, tool use, and multilingual translation across eight languages. It consists of three main elements: Loop components for real-time collaboration, Loop pages for organizing components, and Loop workspaces for shared editing.

The model will be available on cloud platforms, including Google Cloud, AWS , and Snowflake. It also partnered with Accenture to enable its customers to build custom LLM models with the new model.

Analyst QuickTake: Llama 3.1 405 billion is claimed to rival closed models like GPT-4, GPT-4o, and Claude 3.5 Sonnet on various tasks. This may be seen as a step forward in democratizing access to AI technology and enabling global collaboration and innovation.