Meta introduces largest Llama 3 AI model with enhanced language and math capabilities

Meta Platforms has unveiled its largest version of the Llama 3 AI model, boasting impressive multilingual capabilities and performance metrics that challenge paid models from competitors like OpenAI. The new model can converse in eight languages, write better computer code, and solve complex math problems thanks to its 405 billion parameters. That makes it significantly more powerful than its predecessor, though it still needs to catch up to OpenAI’s GPT-4, which has one trillion parameters, and Amazon’s upcoming two trillion parameter model.

CEO Mark Zuckerberg has high expectations for Llama 3, predicting it will surpass proprietary competitors by next year. Meta’s AI chatbot, powered by these models, is on track to become the most popular AI assistant by the end of this year, already used by hundreds of millions. The release comes amidst a competitive push among tech companies to demonstrate the value of their advanced AI models in solving complex reasoning tasks, justifying the significant investments made.

Meta is also releasing updated versions of its lighter-weight 8 billion and 70 billion parameter Llama 3 models. All versions are multilingual and can handle larger user requests, enhancing their ability to generate computer code. Meta’s head of generative AI, Ahmad Al-Dahle, highlighted improvements in solving math problems by using AI to generate training data. By offering Llama models largely free of charge, Meta aims to foster innovation, reduce dependence on competitors, and increase engagement on its social networks, despite some investor concerns over the costs involved.