Tag
Google DeepMind released the Gemma 4 series of open-weight models, covering four sizes from 2B to 31B, supporting 128K–256K context, reasoning, and function calling, under Apache 2.0 license, and equipped with ADK framework for autonomous agent capabilities.