Google’s Gemma AI Models Exceed 150 Million Downloads, Challenging Llama

Google’s Gemma, a family of open AI models, has achieved a significant milestone, surpassing 150 million downloads. This announcement was made by Omar Sanseviero, a developer relations engineer at Google DeepMind, on X (formerly Twitter). According to Sanseviero, the developer community has created over 70,000 variants of Gemma on the Hugging Face platform.

Gemma just passed 150 million downloads and over 70k variants on Hugging Face

What would you like to see in the next Gemma versions?

— Omar Sanseviero (@osanseviero) May 11, 2025

Launched in February 2024, Gemma was designed as a competitor to other open-source model families, notably Meta’s Llama. Recent iterations of Gemma are multimodal, enabling them to process both images and text, and they support over 100 languages. Google has also tailored Gemma for specific applications, including drug discovery and emotion recognition.

While 150 million downloads is an impressive figure within a year, Gemma still lags behind Meta’s Llama, which exceeded 1.2 billion downloads as of late April. This difference highlights the varying levels of adoption within the open-source AI model landscape. The growth trajectory of Gemma underscores the increasing interest and adoption of open AI models among developers and organizations.

However, both Gemma and Llama have faced scrutiny regarding their licensing terms. Some developers have raised concerns that the custom, non-standard licenses associated with these models introduce commercial risks, potentially hindering their widespread use in commercial applications. This issue of licensing is a crucial aspect in the open-source AI space, balancing accessibility with commercial viability.

Despite the licensing concerns, Gemma’s achievement of over 150 million downloads signals strong interest and growing adoption within the AI community. As Google continues to develop and refine Gemma, it will be interesting to observe how it competes with other leading open AI models and addresses the challenges related to licensing and commercial use.

The widespread use of these models underscores the growing importance of open source AI and its impact on various industries. Continuous development and community feedback will likely shape the future capabilities and applications of the AI Model.