Meta's Llama series of AI models has become the fastest-growing open-source AI model family with 350 million downloads globally from Hugging Face. This has come along with 20 million downloads in the past month alone, pointing to an increase in appeal of open-source.
The models are not only gaining traction on community platforms like Hugging Face but are also being extensively used through Meta's cloud partners. These include major players such as Amazon Web Services (AWS), Microsoft Azure, Google Cloud and IBM Watsonx, among others.
Meta reported that hosted Llama usage, measured by token volume, more than doubled from May to July 2024 following the release of Llama 3.1. In some cases, the token volume grew tenfold from January to July 2024 across leading cloud service providers.
“Llama has profoundly impacted the advancement of state-of-the-art AI. The floodgates are now open for every enterprise and industry to build and deploy custom Llama supermodels using Nvidia AI Foundry, which offers the broadest support for Llama 3.1 models across training, optimization, and inference. It’s incredible to witness the rapid pace of adoption in just the past month,” said Jensen Huang, Founder & CEO of Nvidia.
The rapid uptake of Llama 3.1 has reflected in the adoption rates by Databricks as well, which described it as their fastest-adopted and best-selling open-source model ever. Industry leaders are praising the model for its ability to bridge the quality gap between open-source and proprietary AI models, with many companies expressing interest in becoming future integration partners. Notable among these are Wipro, Cerebras and Lambda.
The enthusiasm extends beyond just cloud services. Groq, a key partner in the AI space, has reported that their capacity for Llama is rapidly being consumed, with the demand outstripping their current ability to supply it.
“To date Groq has provided over 400,000 developers with 5 billion free tokens daily, using the Llama suite of models and our LPU Inference. It’s a very exciting time and we’re proud to be a part of that momentum. We can’t add capacity fast enough for Llama. If we 10x’d the deployed capacity it would be consumed in under 36 hours,” said Jonathan Ross, Founder & CEO, Groq.
With over 60,000 derivative models developed on Hugging Face, Llama has enabled a vibrant community of developers fine-tuning the technology for diverse use cases. Major enterprises such as AT&T, DoorDash, Goldman Sachs, and Zoom are among those leveraging Llama's capabilities.