Google has released Gemma, a new open-source large language model (LLM) designed to democratize access to advanced AI. The model, built on the same research as Gemini, is intended for developers and researchers seeking customizable, efficient AI tools.
Gemma comes in two sizes: 2B and 7B parameters, catering to different hardware constraints. Google emphasizes that Gemma is lightweight, enabling deployment on laptops and small cloud instances. The company has also released a toolkit for safety and responsible AI use.
This move aligns with Google's broader strategy to compete in the open-source AI space, following similar releases from Meta and other tech giants. Gemma is available under a permissive license, allowing modification and redistribution.
"We want to make it easier for the community to build with Google's models," a spokesperson said.
The release includes pre-trained and instruction-tuned versions, supporting multiple languages and tasks like summarization, question answering, and code generation. Developers can access Gemma via Google's Vertex AI and other platforms.