Google Gemma 2

Google's open-source family of lightweight language models for diverse deployments

AI Infrastructure & MLOpsFree

Gemma 2 is Google DeepMind's second generation of open-weight language models designed to provide high-quality AI capabilities in compact, efficient packages that can run on consumer hardware. The Gemma 2 family includes 2B, 9B, and 27B parameter models that outperform much larger models from earlier generations, reflecting advances in training data quality and architecture. Released freely for research and commercial use, Gemma 2 is available through HuggingFace and Google AI Studio. Developers building local AI applications, researchers, and organizations needing deployable open models use Gemma 2.

Key Features

• ✓Multiple sizes
• ✓High efficiency
• ✓Open weights
• ✓Consumer hardware compatible
• ✓Strong benchmarks