Introducing Gemma 3 270M: The compact model for hyper-efficient AI

In three linesGoogle DeepMind releases Gemma 3 270M, a compact 270-million parameter model designed for hyper-efficient AI. This addition expands the Gemma 3 toolkit with a lightweight option optimized for resource-constrained deployments.

Summary generated by Claude — human-verified