Google's Gemma-4-31B Models Hit AWS SageMaker JumpStart

Google DeepMind's Gemma-4-31B-it-assistant and NVIDIA's quantized Gemma-4-31B-IT-NVFP4 are now deployable on Amazon SageMaker JumpStart, giving AWS customers access to a powerhouse 31B model that ranks #3 on the Arena AI leaderboard while handling multimodal inputs, coding, and agentic workflows. The NVFP4 variant cuts memory usage by 68% and boosts inference speed 2.5x—perfect for production deployments without breaking the bank.