Google has launched Gemma 3 as its latest open model for developers, highlighting it as the “world’s best single accelerator model.”
An American tech giant on Wednesday, March 12, 2025, announced Gemma 3, which is Google’s family of open models with over 100 million downloads in the past year and 60,000 Gemma variants in what’s dubbed the “Gemmaverse.”
In a blog post, the company said that Gemma 3 is built from the same research and technology that powers the Gemini 2.0 models and it’s available in 1B, 4B, 12B, and 27B sizes.
Related: Gemini updates Gmail with new button for automatic Google Calendar events
Google boosts “advanced text and visual reasoning capabilities” to “analyse images, text, and short videos” on the 4B+ sizes.
Gemma 3 supports function calling and structured output to help you automate tasks and build agentic experiences.
Moreover, it introduces official quantised versions, reducing the model size and computational requirements while maintaining high accuracy.
Gemma 3’s vision encoder is also updated with support for high-resolution and non-square images, while the new ShieldGemma 2 image safety classifier is available for use to filter both image input and output for content classified as sexually explicit, dangerous, or violent.
Related: Google Chrome updates extension policy to limit deceptive affiliate practices
Users can easily try it now in Google AI Studio, while model downloads are available through Kaggle or Hugging Face.
