- Added NVIDIA Qwen3.6-35B-A3B-NVFP4 model to SageMaker JumpStart, offering agentic coding, multimodal reasoning, and long-context MoE capabilities with reduced memory footprint.
- Added Alibaba Wan2.1-T2V-1.3B-Diffusers model for lightweight text‑to‑video generation, running on consumer‑grade hardware with minimal VRAM.
- Both models can be deployed via the JumpStart console or SageMaker Python SDK with just a few clicks.