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

New Models Available in Amazon SageMaker JumpStart

AWS has introduced Qwen3-VL-Embedding-2B and Qwen3-Reranker-4B in Amazon SageMaker JumpStart, enhancing the range of foundation models for information retrieval and cross-modal understanding. These models are designed to build comprehensive search pipelines on AWS infrastructure.

Qwen3-VL-Embedding-2B

• Inputs: Text, images, screenshots, videos, and mixed modalities
• Outputs: Semantically rich vectors capturing visual and textual information
• Tasks: Image-text retrieval, video-text matching, visual question answering, multimodal content clustering
• Languages: Over 30 languages

Qwen3-Reranker-4B

• Inputs: Query and document pairs
• Outputs: Relevance scores to refine retrieval results
• Tasks: Text retrieval, code retrieval, text classification, text clustering, bitext mining
• Languages: Over 100 languages

What to do

• Deploy models with a few clicks in SageMaker Studio
• Use the SageMaker Python SDK for deployment
• Refer to the Amazon SageMaker JumpStart documentation for more information

Source: AWS release notes

If you need further guidance on AWS, our experts are available at AWS@westloop.io. You may also reach us by submitting the Contact Us form.