Google's Gemma AI Models Reach 150 Million Downloads

May 12, 2025
Google's Gemma AI models have surpassed 150 million downloads, with over 70,000 variants created on Hugging Face. Despite this milestone, Gemma trails behind Meta's Llama, which has over 1.2 billion downloads.
Google's Gemma AI Models Reach 150 Million Downloads

Google has announced that its Gemma AI models have surpassed 150 million downloads. This milestone was shared by Omar Sanseviero, a developer relations engineer at Google DeepMind, on X, where he also noted that developers have created more than 70,000 variants of Gemma on the AI development platform Hugging Face.

Launched in February 2024, Gemma was designed to compete with other open model families, such as Meta's Llama. The latest versions of Gemma are multimodal, capable of processing both images and text, and support over 100 languages. Google has also developed specialized versions of Gemma for specific applications, including drug discovery.

Despite the impressive download numbers, Gemma still lags behind its rival, Llama, which has achieved over 1.2 billion downloads. Both Gemma and Llama have faced criticism for their custom, non-standard licensing terms, which some developers argue complicate commercial use.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Daily AI Brief.

Also, consider following us on social media:

AI Brief AI Brief (X)
Free newsletter

Daily AI Brief

Daily report covering major AI developments and industry news, with both top stories and complete market updates

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.