Gemma 4 models are now available in Amazon SageMaker JumpStart

AWS Announces New Gemma Models in SageMaker JumpStart

AWS has introduced three new Gemma models in Amazon SageMaker JumpStart: Gemma 4 E4B, Gemma 4 26B-A4B, and Gemma 4 31B. These models offer multimodal capabilities, including configurable reasoning, native function calling, and multilingual support across 140+ languages.

Key Features

• Thinking: Built-in reasoning mode for step-by-step answers
• Image Understanding: Object detection, document parsing, screen understanding, chart comprehension, OCR, and handwriting recognition
• Video Understanding: Analyze video content by processing sequences of frames
• Interleaved Multimodal Input: Mix text and images in any order within a single prompt
• Function Calling: Native support for structured tool use
• Coding: Code generation, completion, and correction
• Multilingual: Support for 35+ languages, pre-trained on 140+ languages
• Audio Input: Gemma 4 E4B supports automatic speech recognition (ASR) and speech-to-translated-text translation

What to do

• Deploy models via SageMaker Studio or SageMaker Python SDK
• Refer to the Amazon SageMaker JumpStart documentation for deployment details

Source: AWS release notes

If you need further guidance on AWS, our experts are available at AWS@westloop.io. You may also reach us by submitting the Contact Us form.