Gemma 4 models are now available in Amazon SageMaker JumpStart

AWS announced the availability of three new Gemma 4 foundation models in Amazon SageMaker JumpStart: Gemma 4 E4B, Gemma 4 26B-A4B, and Gemma 4 31B. These instruction-tuned models from Google DeepMind introduce multimodal capabilities that combine text, image, and video processing with configurable reasoning, native function calling, and multilingual support across more than 140 languages. The models can be deployed with just a few clicks through SageMaker JumpStart, enabling AWS customers to build sophisticated AI applications on the cloud platform. The Gemma 4 models bring several advanced capabilities including built-in step-by-step reasoning, comprehensive image understanding for document parsing and OCR, video content analysis through frame processing, and interleaved multimodal input that allows mixing text and images within single prompts. Additionally, the models support native function calling for agentic workflows, code generation and correction, and out-of-the-box multilingual capabilities. The Gemma 4 E4B variant includes unique audio input support for automatic speech recognition and speech-to-translated-text functionality across multiple languages.