Deploy Qwen models with Amazon Bedrock Custom Model Import

Amazon Bedrock Custom Model Import now supports Qwen models, allowing organizations to deploy Qwen 2, 2.5 Coder, Qwen 2.5 VL, and QwQ 32B models in a fully managed, serverless environment. The service enables importing custom model weights for Qwen2, Qwen2_VL, and Qwen2_5_VL architectures without managing infrastructure. The post demonstrates two use cases: deploying Qwen2.5-Coder-7B-Instruct as a coding assistant and using Qwen2.5-VL for image understanding tasks. The process involves downloading models from Hugging Face, uploading to S3, and importing through the Bedrock console or API. Pricing is based on active model copies and duration, with automatic scaling and pay-per-use billing in 5-minute increments.