Gemma 4 models now available on Amazon Bedrock

This article announces the availability of Google DeepMind's Gemma 4 family of open-weight models on Amazon Bedrock for building generative AI applications.

• Three Gemma 4 variants available: 31B, 26B-A4B, and E2B with different architectures
• Supports reasoning, multimodal understanding, agentic, and software engineering workflows
• Features native function calling, 35+ language support, and multimodal input capabilities
• Gemma 4 31B optimized for reasoning and coding with 256K-token context window
• Gemma 4 26B-A4B targets cost and latency-sensitive applications
• Gemma 4 E2B designed for low-latency interactive use cases
• Available in US East (N. Virginia), US East (Ohio), US West (Oregon), and Europe (Frankfurt)

Gemma 4 models on Bedrock enable developers to build reliable AI applications using open-source models with improved tool calling, structured output, and reasoning capabilities.