Meta releases early versions of its Llama 3 AI model

Meta Platforms has launched early versions of its latest large language model, Llama 3, and an image generator aimed at rivaling OpenAI in the generative AI market. These models will be integrated into Meta’s virtual assistant, Meta AI, which promises superior performance in reasoning, coding, and creative writing compared to competitors like Google and Mistral AI. Meta plans to feature Meta AI prominently in its Facebook, Instagram, WhatsApp, and Messenger apps, as well as on a standalone website, directly competing with OpenAI’s ChatGPT. The company aims to challenge OpenAI’s leading position by providing powerful AI tools for free, despite concerns about safety and misuse. Llama 3, trained with both text and images, will initially output only text, with plans for multimodal capabilities in future versions. Meta also announced partnerships with Google for real-time search results and plans to expand Meta AI to more international markets. CEO Mark Zuckerberg highlighted Llama 3’s performance benchmarks, indicating a growing performance gap between free and proprietary models. Despite past issues with context understanding, Meta claims to have improved Llama 3’s capabilities through the use of high-quality data.

Source – CGTN