Google’s Gemma AI Models Cross 150 Million Downloads
Google’s Gemma AI models have surpassed a major milestone, recording over 150 million downloads, according to a weekend announcement by Omar Sanseviero, a developer relations engineer at Google DeepMind. The figures underscore growing interest in Google’s open model family but also highlight the intense competition it faces in the AI model space.
Gemma’s Rapid Rise Since Launch
Launched in February 2024, Gemma was positioned as Google’s answer to other open-access model families like Meta’s Llama. In just over a year, developers have built more than 70,000 custom versions of Gemma on platforms like Hugging Face, reflecting a healthy and active developer ecosystem.
Gemma models are multimodal, supporting both text and image inputs, and are compatible with over 100 languages. This flexibility has made them appealing for a range of tasks, from general-purpose AI assistants to specialized use cases like drug discovery.
Also read: China’s Manus AI Opens Free Registration Following $75M Funding
Google has also introduced fine-tuned versions of Gemma tailored for specific industries, boosting its utility across enterprise applications.
Still Trailing Meta’s Llama in Popularity
Despite strong adoption, Gemma still lags behind Meta’s Llama, which crossed 1.2 billion downloads by late April 2025. Meta’s head start and aggressive distribution strategy have made Llama a dominant player in the open model landscape.
Both Gemma and Llama have been subject to criticism over non-standard licensing terms. While marketed as “open,” these models include clauses that can restrict commercial usage, creating uncertainty for companies looking to deploy them in real-world products.
The Bigger Picture
Google’s progress with Gemma signals a broader trend: the race to lead in open foundation models is far from over. While Meta holds the current lead, Google is steadily building momentum with enhanced multimodal capabilities and growing developer adoption.
As demand for adaptable, open-access models continues to surge, Google’s challenge will be to combine openness, scalability, and enterprise-ready features—without limiting commercial flexibility.
