LocateAnything-3B, Qwen-AgentWorld-35B-A3B, and Qwen3.5-122B-A10B models now available on Amazon SageMaker JumpStart
AWS has added three new foundation models to Amazon SageMaker JumpStart: NVIDIA’s LocateAnything-3B, Qwen’s Qwen-AgentWorld-35B-A3B, and Qwen’s Qwen3.5-122B-A10B, expanding AI deployment options for customers.
Amazon SageMaker JumpStart now hosts NVIDIA’s LocateAnything-3B model, which specializes in visual grounding and object localization from natural language instructions. The model uses a Parallel Box Decoding framework to process bounding boxes and points in a single step, improving geometric coherence and enabling parallel execution. It supports precise object localization, dense detection, and point-based localization across enterprise and physical AI applications. Customers can deploy it directly through the SageMaker JumpStart model catalog or via the SageMaker Python SDK.
Qwen’s Qwen-AgentWorld-35B-A3B has been added to SageMaker JumpStart, offering agent environment simulation across seven interaction domains. These include tool calling, search, terminal, software engineering, Android, web, and OS interaction. The model predicts next environment states using long chain-of-thought reasoning and was trained on over 10 million real-world interaction trajectories. It is the first language world model to cover all seven domains within a single architecture, providing a unified solution for agent-based AI workflows.
Qwen’s Qwen3.5-122B-A10B is now available on SageMaker JumpStart, delivering high-performance multimodal reasoning with production-friendly efficiency. The model features 122 billion total parameters but activates only 10 billion per token through a hybrid architecture combining Gated Delta Networks with sparse Mixture-of-Experts (256 experts). It supports a native 262K context window and is optimized for reasoning, coding, agent tasks, and visual understanding with minimal latency overhead.
AWS customers can deploy any of these three models through Amazon SageMaker JumpStart with minimal setup. The models can be accessed via the SageMaker JumpStart model catalog in the SageMaker console or deployed programmatically using the SageMaker Python SDK. AWS provides documentation to guide users through the deployment and integration of foundation models in SageMaker JumpStart for their specific AI use cases.