Amazon SageMaker AI now supports serverless model customization for Qwen3.5 models
Amazon Web Services has expanded its SageMaker AI platform to support serverless model customization for Alibaba Cloud's Qwen3.5 family of large language models. The new capability allows developers to fine-tune Qwen3.5 models with 4 billion, 9 billion, and 27 billion parameters using supervised fine-tuning (SFT) and reinforcement fine-tuning (RFT) techniques. Previously, organizations could only deploy the base Qwen3.5 models on SageMaker AI without customization options. The serverless approach eliminates the need for infrastructure management, with AWS handling all provisioning and training orchestration while customers pay only for actual usage. Organizations can now adapt these open-weight models using proprietary data to improve domain-specific accuracy, align outputs with organizational standards, or enhance performance on specialized tasks. The service is currently available in four AWS regions: US East (N. Virginia), US West (Oregon), Asia Pacific (Tokyo), and EU (Ireland), with access through Amazon SageMaker Studio or the SageMaker Python SDK.
Why It Matters
This expansion demonstrates AWS's strategy to support popular open-source AI models alongside proprietary offerings, potentially reducing vendor lock-in concerns for enterprises. The serverless model customization capability addresses a key enterprise AI challenge - making foundation models domain-specific without requiring extensive ML infrastructure expertise. By supporting Alibaba's Qwen3.5 models, AWS is also signaling its willingness to integrate Chinese-developed AI technologies despite ongoing geopolitical tensions in the tech sector.
This summary is generated using AI analysis of the original press release. Always refer to the original source for complete details.