aws AWS What's New ·

NVIDIA and Qwen AI Models Now on Amazon SageMaker JumpStart

aiawsgaengineeraws-sagemaker
feature

Amazon SageMaker JumpStart now offers NVIDIA's LocateAnything-3B and Qwen's Qwen-AgentWorld-35B-A3B and Qwen3.5-122B-A10B models. These models expand the portfolio of foundation models for AWS customers, providing specialized capabilities in visual grounding, agent environment simulation, and multimodal reasoning. LocateAnything-3B uses a Parallel Box Decoding framework for precise object localization, while Qwen-AgentWorld-35B-A3B simulates agent environments across seven interaction domains. Qwen3.5-122B-A10B delivers efficient multimodal reasoning with a hybrid architecture and 262K context window, all deployable via SageMaker JumpStart.

  • NVIDIA LocateAnything-3B for Visual Grounding
  • Qwen-AgentWorld-35B-A3B for Agent Environment Simulation
  • Qwen3.5-122B-A10B for High-Performance Multimodal Reasoning
Features (3)
  • NVIDIA LocateAnything-3B for Visual Grounding

    This model is optimized for fast, high-quality visual grounding and object localization from natural language instructions. It uses a Parallel Box Decoding (PBD) framework to decode bounding boxes and points as atomic units, enabling precise localization across diverse domains in both Enterprise Intelligence and Physical AI applications.

  • Qwen-AgentWorld-35B-A3B for Agent Environment Simulation

    This model excels in simulating agent environments across seven interaction domains, including tool calling, search, and software engineering. It is the first language world model to cover all seven domains, predicting next environment states via long chain-of-thought reasoning trained on over 10 million real-world interaction trajectories.

  • Qwen3.5-122B-A10B for High-Performance Multimodal Reasoning

    This model provides high-performance multimodal reasoning with production-friendly efficiency. It features 122B parameters with sparse Mixture-of-Experts (256 experts), delivering strong reasoning, coding, agents, and visual understanding performance with a native 262K context window and minimal latency overhead.

Read the original announcement →

https://aws.amazon.com/about-aws/whats-new/2026/01/locateAnything-3B-qwen-agentworld-35B-A3B-qwen3.5-122B-A10B-on-sagemaker-jumpstart/

Related releases