Gemma 4 includes three model variants—31B, 26B-A4B, and E2B—covering dense and mixture-of-experts architectures with extensive multimodal capabilities across text, images, video, and audio. The models support 35+ languages and feature long context windows, designed for reasoning, coding, and low-latency interactive applications.
Available in AWS GovCloud (US-West), Gemma 4 runs on Amazon Bedrock’s new inference platform optimized for pricing and performance. Features include native function calling, structured outputs, response streaming, and tool calling, enabling customers to build reliable, flexible generative AI applications in secure and compliance-conscious environments.
Source & further reading
Gemma 4 models are now available on Amazon Bedrock in AWS GovCloud (US-West) ↗Source published 2026-07-30. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.