Gemma 3

About Gemma 3

Gemma 3, developed by Google DeepMind, is an open-source model family available in sizes from 1B to 27B parameters, built from research underlying the Gemini program. The model supports multimodal inputs including text and images, along with over 140 languages, making it one of the most versatile open-source models available. Gemma 3 achieves competitive performance with much larger models through efficient architecture design and training techniques developed for the Gemini model line. Its compact sizes enable deployment on consumer hardware from laptops to mobile devices, democratizing access to capable multimodal AI. Free and open-source under Google's permissive license, it supports commercial use and fine-tuning. The model represents Google's strategy of releasing capable open-source models derived from its proprietary Gemini research, building developer ecosystem engagement while maintaining the commercial advantage of its larger models.