Meta has been talking about releasing a great open-source AI model since April and now on July 23, it has finally released its latest AI model. Meta Llama 3.1 is the company’s first model for the AI ​​industry. The company claims that this model is the greatest open-source AI model ever built and has already outperformed top ranked competitors. In the blogpost, Meta said that the latest model has outperformed GPT 4o and Anthropic’s Cloud 3.5 Sonnet on several benchmarks. Last year, Llama 2 was only on par with the older generation model behind Frontier. This year, Llama 3 is competitive with the most advanced models and is leading in some areas.

What is Llama 3.1?

Llama 3.1 is the most complex AI model built by Meta to date. It is interesting to note that the recent Meta AI release is based on Llama 3, and it is quite fun to use. In the blog post, Meta CEO Mark Zuckerberg said that just like Linux surpassed proprietary operating systems, open-source AI models are advancing rapidly and are set to surpass proprietary operating systems, indicating the direction of Linux becoming the dominant force in powering modern devices and infrastructure.

He draws parallels between Meta’s investments in open-source AI and its previous Open Compute project, which yielded significant cost savings by leveraging collaborative innovation. He expects similar results with AI, where open-source development will drive progress and efficiency, mirroring the Open Compute project’s success in data center design and operations.

He added, “I believe the Llama 3.1 release will be a turning point in the industry, where most developers begin to use open source as their primary focus, and I expect this approach to grow from here on out.”

These models will be available on all major clouds including AWS, Azure, Google, Oracle and others. Scale.AI, Dell, Deloitte and other companies are set to help enterprises adopt Llama and train custom models with their own data.

It is not surprising that Meta has not mentioned anything when it comes to the training of the model. It has not revealed the actual source data used to train Llama 3.1.

Meta AI is now multilingual

Meta is expanding the reach of its llama-based AI assistant to additional countries and languages, as well as introducing a new feature that enables image generation based on a person’s unique appearance. Meta has also announced that the MetaAI assistant in Meta apps is now going multilingual. Meta AI is now available in 22 countries, with the newest being Argentina, Chile, Colombia, Ecuador, Mexico, Peru, and Cameroon.

Users can interact with Meta AI on WhatsApp, Instagram, Messenger, and Facebook in new languages: French, German, Hindi, Hindi-Roman script, Italian, Portuguese, and Spanish, with more to come.