NVIDIA Nemotron 3.5 Super VL is now available in Amazon SageMaker JumpStart

What happened

NVIDIA's Nemotron 3.5 Super VL, a 120-billion-parameter vision-language model, is now available for deployment through Amazon SageMaker JumpStart. Users can set up the model in SageMaker Studio or with the SageMaker Python SDK. The model can process images, video, and text, and it generates text outputs. Two variants are offered: NVFP4, optimized for inference, and BF16, intended for customization.

Key details

• Nemotron 3.5 Super VL has 120 billion total parameters, with 12 billion active per token.
• The model supports context lengths up to 1 million tokens, though JumpStart configurations use 520 to 15,600 tokens.
• NVFP4 checkpoint is about 75 GB and runs with 4-bit weights and activations on Blackwell GPUs; BF16 is about 249 GB.
• Supported input types are image, video, and text; output is text only.
• The model is licensed under the OpenMDW License Agreement, version 1.1.

Why it matters

Organizations using Amazon SageMaker JumpStart can now deploy a large-scale vision-language model capable of handling complex multimodal tasks, such as document analysis, GUI understanding, and video question answering. This provides businesses and developers with access to advanced AI capabilities for processing and reasoning over images, videos, and text, without needing to manage the infrastructure for such a large model. The availability of both optimized and customizable variants allows users to choose between faster inference or further model tuning for specific needs.