Meta has released its largest open-source AI model, Llamas 3.1.405 b. Meta claims that this model has 405 billion parameters, making it one of the most advanced and capable models in the open-source AI space.
The model will be available for free, enabling wider access to the developer community. It was trained using Nvidia's H100 GPU (Graphics Processing Unit) and leverages advanced training and development techniques.
Meta claims that the model will be able to beat OpenAI's GPT-4o and Anthropik's Cloud 3.5 Sonnet in several benchmarks. Developers can run it on their own infrastructure using Llama 3.1.405 B for about half the cost of using other proprietary models such as GPT-4o for user-facing, offline inference applications.
In an open letter to Meta CEOs, Zuckerberg stressed the importance of open source for a positive AI future. Similar to how Linux has evolved into the industry standard foundation for cloud computing and the operating systems that most mobile devices run on, open-source AI models are advancing faster than proprietary models.
"I believe open source is necessary for a positive AI future. AI has more potential than any other modern technology to increase human productivity, creativity, and quality of life, and to accelerate economic growth while unlocking advances in medical and scientific research.
"Open source will ensure that more people around the world have access to the benefits and opportunities of AI, that power isn't concentrated in the hands of a few companies, and that technology can be more evenly and securely deployed across society," Zuckerberg said.
the company's scale. AI is actively building partnerships with companies like Dell and Deloitte, which aims to help companies adopt Llama and train custom models using their own data. Llama 3.1.405 B is available on cloud platforms such as AWS, Azure, and Google Cloud.
"With past Llama models, Meta developed them themselves and then released them, but wasn't too focused on building a broader ecosystem. We take a different approach with this release. We're building teams internally to enable as many developers and partners as possible to use Llama, and we're actively building partnerships with more companies in the ecosystem to deliver unique functionality to their customers as well.
Aiming to mitigate risks before release, he said Meta's security process includes rigorous testing and red teaming to identify risks from their models. Since the models are open source, others can also run their own tests.
Zuckerberg compared Meta's investment in open-source AI to its previous Open Compute project and claimed it helped the company make billions.
"Meta has a long history of open source projects and successes. We've saved billions of dollars by using the Open Compute Project to roll out our server, network, and data center designs, and standardizing supply chains in our designs. "He said.
