OFICIAL AWS What's New

Qwen3 embedding and reranking models for retrieval are now available in Amazon SageMaker JumpStart

What happened
Based on AWS What's New · Jul 13, 2026

AWS has made Qwen3 embedding and reranking models available in SageMaker JumpStart for retrieval workflows, enabling multimodal and multilingual search pipelines.

🚩 Report a correction ↗ Open source
Key points
·
Today, AWS details the availability of Qwen3-VL-Embedding-2B and Qwen3-Reranker-4B in Amazon SageMaker JumpStart, expanding the portfolio of foundation models available to AWS customers.
·
These models from Qwen are designed for information retrieval and cross-modal understanding, enabling customers to build comprehensive search pipelines on AWS infrastructure.
·
The two models are typically used in tandem: the embedding model performs efficient initial recall, while the reranker refines results in a subsequent re-ranking stage.
·
It delivers performance across diverse multimodal tasks such as image-text retrieval, video-text matching, visual question answering, and multimodal content clustering, with support for over 30 languages.
Key numbers
·
Qwen3-Reranker-4B evaluates query-document pairs to produce relevance scores, refining retrieval results for applications including text retrieval, code retrieval, and text classification.
·
It supports over 100 languages and allows user-defined instructions to tailor performance for specific tasks or scenarios.
·
AWS has made Qwen3 embedding and reranking models available in SageMaker JumpStart for retrieval workflows, enabling multimodal and multilingual search pipelines.

Amazon Web Services announced the availability of Qwen3-VL-Embedding-2B and Qwen3-Reranker-4B in Amazon SageMaker JumpStart, expanding the foundation model options for AWS customers. The models are designed for information retrieval and cross-modal understanding, allowing users to construct search pipelines on AWS infrastructure. The embedding model processes diverse inputs such as text, images, videos, and mixed modalities, generating semantically rich vectors for tasks like image-text retrieval and multimodal clustering across over 30 languages.

Qwen3-Reranker-4B evaluates query-document pairs to produce relevance scores, refining retrieval results for applications including text retrieval, code retrieval, and text classification. It supports over 100 languages and allows user-defined instructions to tailor performance for specific tasks or scenarios. The reranker operates as a follow-up to the embedding model, enhancing the precision of search outcomes in retrieval workflows.

Customers can deploy either model in SageMaker JumpStart with minimal setup, enabling rapid integration into existing AI use cases. The deployment process can be initiated through SageMaker Studio’s Models section or via the SageMaker Python SDK, providing flexibility for different workflow preferences.

For detailed guidance on deploying and using these foundation models, AWS directs users to the Amazon SageMaker JumpStart documentation. The models are now accessible to AWS account holders for immediate implementation in retrieval and multimodal applications.

Original source → Deals on Clipraptor.com →