Meta Platforms Inc. today revealed that its Llama generative artificial models are rapidly growing in popularity, with downloads of the models increasing by 10 times from the same time last year and are quickly approaching 350 million downloads.
The figures were revealed in an upbeat midyear update from the company behind services such as Facebook, Instagram and WhatsApp, which noted that Llama is now the leading open-source generative AI model family.
Chief Executive Officer Mark Zuckerberg’s decision to offer the models open source has always been interesting or, some might argue, a brave move by Meta. But given that AI isn’t the company’s core business — albeit with plenty of cross-over opportunities — the decision to open-source the Meta designed and trained models is starting to see success and the figures are impressive.
Since the launch of Llama 3 in April and Llama 3.1 in July, Meta reports that its models were downloaded more than 20 million times in the last month and that token volume across major cloud service provider partners has more than doubled in the three months from May to July. Monthly usage, by token volume, has also increased by 10 times from January to July.
The latest Llama 3.1 model excels at capabilities such as general knowledge, math, tool use and multilingual translation. The model also saw added support for eight new languages, including French, German, Hindi, Italian, Portuguese and Spanish, with more on the way.
Meta attributes the success to the models being open source.
“By making our Llama models openly available we’ve seen a vibrant and diverse AI ecosystem come to life where developers have more choice and capability than ever before,” Meta stated in the report. “The innovation has been broad and rapid, from startups pushing new boundaries to enterprises of all sizes using Llama to build on-premises or through a cloud service provider.”
Meta revealed in the report that there are now more than 60,000 derivative models on Hugging Face, a company and open-source community that provides tools, libraries and models for machine learning.
The models are also being embraced by corporations, with Llama model users now, including AT&T Inc., DoorDash Inc., Goldman Sachs Group Inc., Niantic Inc., Nomura Holdings Inc., Shopify Inc., Spotify Technology S.A. and Zoom Video Communications Inc. In addition, Infosys Ltd. and KPMG International Ltd. are also said to be using Llama models internally.
“We’re excited by the growth of the Llama community and encouraged knowing we’re building the most advanced large language models, open-sourced for the world today,” the report concludes.