Gemma 2

About Gemma 2

Gemma 2, developed by Google DeepMind, is an open-source language model family available in 2B, 9B, and 27B parameter sizes with an 8K token context window. The model family brings research-grade capabilities from the Gemini program to the open-source community, performing well on reasoning, coding, and general knowledge tasks relative to its size class. Gemma 2 can be fine-tuned for specific domains and runs efficiently on consumer GPUs, making it accessible for independent researchers and small organizations. Its permissive license allows commercial use and modification. Priced at zero cost as a fully open-source release, it has become widely adopted for academic experiments in alignment, efficiency, and domain adaptation. Gemma 2 ranks #26 on the Chatbot Arena leaderboard, reflecting solid performance for an open-weight model.