Meta Platforms Inc. is making its artificial intelligence large language model, Llama 2, available for commercial use through partnerships with major cloud providers including Microsoft Corp.
Meta isn’t charging for access or usage of the model it developed, the company said. Instead, by opening up the technology to other companies, Meta says it will benefit from improvements that can be made when more developers use, stress test and identify problems with it.
Making the large language model, or LLM, more widely accessible, also establishes Meta alongside other tech giants as having a key contribution to the AI arms race. Meta is spending record amounts on AI infrastructure, and Chief Executive Officer Mark Zuckerberg has said incorporating AI improvements into all the company’s products and algorithms is a priority.
Meta took on the cost of training the models. Cloud providers including Microsoft, Amazon.com Inc. and Hugging Face are hosting the tools and providing the computing power to run them.
Meta said it had no comment on whether Microsoft will charge for access to Llama 2 via its cloud services and on Windows. An Amazon spokesperson said the company isn’t charging for access to the model, and customers only pay for usage of SageMaker, its toolkit for machine learning developers. Microsoft and Hugging Face didn’t immediately respond to requests for comment.
The commercial rollout of Llama 2 is the first project to debut out of the company’s generative AI group, a new team assembled in February. To prepare the new model for release, Meta employees and third-parties ran safety tests called “red-teaming exercises.” It’s a separate model from the one Meta uses for its own products.
See also: Without regulator buy-in, scaling AI in financial services will be an uphill battle
“The pace has been incredible — incredibly quick decision making, really tight loops, really focused execution to lead up to this Llama 2 moment,” said Ahmad Al Dahle, vice president of Meta’s generative AI group. “We had over a hundred thousand requests for Llama 1. So we went back to the drawing board, brought our teams together, and made a ton of investments on alignment and research and safety in order to build a model that we felt was was ready.”
The partnership aligns Meta more closely with Microsoft, which has emerged as the clear leader in AI through its investment and technology partnership with ChatGPT creator OpenAI, which charges for access to its model.
Microsoft’s tie-up with OpenAI doesn’t concern Meta, says Al Dahle. He said Llama 2, being open-sourced, can provide the opportunity for small- and medium-sized businesses to more quickly test and use LLMs without owning the servers and computing power to host their own custom model.
Large language models, or LLMs, power text-based chatbots, though their uses aren’t limited to helping AI answer questions. The models are trained by being fed immense amounts of information, usually from the internet, to refine a process for generating responses.
“Meta has a long history of open sourcing our infrastructure and AI work,” Zuckerberg said in a post on Facebook. “This has helped us build better products by driving progress across the industry.” Meta will also be opening Llama 2 access for academic researchers.
Previously, the company developed PyTorch, an AI framework, to be accessible to all and eventually spun it out to be overseen by the nonprofit Linux Foundation.