AWS Expands Amazon Bedrock with Chinese AI Models Qwen3 & DeepSeek-V3.1

Amazon Web Services (AWS) announced that it has expanded its Amazon Bedrock lineup by adding Alibaba’s Qwen3 family of models and DeepSeek-V3.1, both of which are now available globally.

Last month, AWS announced that OpenAI’s open weight models—gpt-oss-120b and gpt-oss-20b—are now available via Amazon Bedrock.

"By using these models in Amazon Bedrock, customers benefit from enterprise-grade security, including data encryption and strict access controls, which help customers maintain data privacy and regulatory compliance," AWS said.

The Qwen3 family introduces four new models capable of handling multilingual tasks—particularly strong in Chinese and English—along with software development, workflow planning, and reasoning. Variants such as Qwen3-Coder are optimized for complex coding tasks and agent-based automation, while others like Qwen3-235B balance efficiency with general reasoning across math and science.

Meanwhile, DeepSeek-V3.1 brings hybrid reasoning capabilities, allowing users to switch between step-by-step “thinking mode” and faster response outputs. With its 685 billion parameters and mixture-of-experts architecture, the model delivers high performance while remaining compute-efficient, excelling in coding, data analysis, and automation.

Companies like CoinMarketCap are already using DeepSeek models in Bedrock to process and analyse massive amounts of crypto market data, turning them into actionable insights.

AWS argues that open-weight models foster transparency and innovation by letting customers adapt and build upon them. With their inclusion in Bedrock across regions from the U.S. to Asia and Europe, AWS is positioning itself as a leading hub for enterprises seeking secure, scalable generative AI tools.

The move strengthens AWS’s portfolio of open weight models, which already includes contributions from Meta, Mistral AI, and OpenAI.