Meta launches Llama 3.1, next-generation AI

A Meta announced last Tuesday (23) the launch of Llama 3.1the latest version of its advanced language model for Generative AI. This model promises to be more efficient than its competitors. GPT-4 Omni and Claude 3.5 Sonnet. Meta partner companies, such as Nvidia, Google Cloud, AWS and Azure can now integrate Llama 3.1 into their applications.

Main features of Llama 3.1

O Llama 3.1 can process up to 128,000 characters or tokens at a timewhich means it can understand and respond to longer prompts more accurately. This is useful for tasks like analyzing detailed reports or spreadsheet data and providing contextualized answers.

In comparative tests, Llama 3.1 outperformed GPT-4 Omni and Claude 3.5 Sonnet in seven different benchmarks. Interested developers can try out Llama 3.1 on official site of model.

Mark Zuckerberg, CEO of Meta, published a carta where he argues that the Open source AIs are the future. He compares this approach to the success of Linux, which has become dominant in cloud computing and mobile devices thanks to its open-source nature.

Zuckerberg believes AIs will follow a similar path, citing the progress of Llama 3 over Llama 2 thanks to this philosophy. He also sees open ecosystems as key to success in the virtual reality market. Below is an excerpt from Mark Zuckerberg’s letter:

“Open source Linux gained popularity initially because it allowed developers to modify the code as they wished and was more accessible, and over time it became more advanced, more secure, and with a more capable ecosystem than any closed Unix. Today, Linux is the foundational standard for cloud computing and the operating system on most mobile devices—and we all benefit from superior products because of it.

I believe AI will develop along similar lines. Today, many technology companies develop closed models. But open source is closing the gap. Last year, Llama 2 was comparable to an older generation of models, before the latest generation. This year, Llama 3 is competitive with most advanced models and leads in some areas. Starting next year, we expect future Llama models to be the most advanced in the industry.“.

Llama 3.1 is not the “Meta AI”

It is important to note that Llama 3.1 is not the same generative AI tool as Meta, which remains unavailable in Brazil following a ban by the ANPD related to the collection of user data. Llama 3.1 is an LLM (Large Language Model), a core technology for the development of artificial intelligences.

With LLMs, developers can create AI tools that leverage the knowledge gained by Llama 3.1.

Speaking of which, Meta’s AI platform recently received updates, expanding its reach to seven more countries and adding support for seven more languages.