Amazon SageMaker JumpStart now offers two new foundation models, expanding AI capabilities for AWS customers. NVIDIA's Qwen3.6-35B-A3B-NVFP4 is optimized for agentic coding and long-context reasoning. This model, a quantized variant of Alibaba's Qwen3.6-35B-A3B, features a Mixture-of-Experts architecture with 35 billion total parameters. It activates only 3 billion parameters per token, supporting a vast 262K-token context window. Quantized to NVFP4, it significantly reduces memory usage while maintaining performance. The model excels at preserving context across conversation turns and multi-step agent tasks. Alibaba's Wan2.1-T2V-1.3B-Diffusers model specializes in text-to-video generation. Built with a diffusion transformer and a Video Variational Autoencoder, it produces high-quality video clips from text prompts. This 1.3 billion parameter model requires minimal VRAM, making it accessible for consumer-grade hardware. It can generate a 5-second 480p video in about four minutes on an RTX 4090. Both models can be easily deployed through SageMaker JumpStart with just a few clicks. Customers can find them in the SageMaker console or deploy them using the SageMaker Python SDK.