Key Points:
- Meta has unveiled its newest open-source model, Llama 3.1 405B.
- This model boasts “state-of-the-art” abilities in areas such as math, translation, and general knowledge.
- Meta also revealed enhancements for Meta AI, which will be accessible in additional countries globally.
Meta’s Latest AI Release: Llama 3.1
Since April, Meta has been discussing the release of a robust open-source AI model. On July 23, it finally introduced its latest AI model, Llama 3.1, marking a significant milestone for the company in the AI industry. Meta claims that this is the largest open-source AI model ever created, outperforming top competitors. According to Meta’s blog post, Llama 3.1 has surpassed GPT-4 and Anthropic’s Claude 3.5 Sonnet on several benchmarks. While Llama 2 was comparable to older models, Llama 3.1 competes with and leads some of the most advanced models available today.
What is Llama 3.1?
Llama 3.1 is Meta’s most sophisticated AI model to date, built on the foundation of Llama 3 and designed to be highly user-friendly. In a blog post, Meta CEO Mark Zuckerberg compared the rise of open-source AI models to the success of Linux, predicting that open-source AI will soon surpass proprietary models, much like Linux has become dominant in powering modern devices and infrastructure.
Zuckerberg highlighted the parallels between Meta’s investment in open-source AI and its previous Open Compute Project, which brought significant cost savings through collaborative innovation. He expects open-source AI development to yield similar advancements and efficiencies. He stated, “I believe the Llama 3.1 release will be a turning point in the industry, where most developers begin to primarily use open source, and I expect that approach to only grow from here.”
Llama 3.1 will be available on major cloud platforms, including AWS, Azure, Google, Oracle, and more. Companies such as Scale.AI, Dell, Deloitte, and others are ready to assist enterprises in adopting Llama and training custom models with their data.
Meta AI Goes Multilingual
Meta is also expanding the reach of its Llama-based AI assistant to more countries and languages. A new feature allows for image generation based on a person’s unique appearance. Meta AI is now available in 22 countries, with recent expansions to Cameroon, Ecuador, Argentina, Chile, Colombia, Peru and Mexico.
Users can interact with Meta AI across WhatsApp, Instagram, Messenger, and Facebook in new languages: French, German, Hindi, Hindi-Romanised Script, Italian, Portuguese, and Spanish, with more languages to follow.
Leave a Reply