Amazon SageMaker AI now supports serverless reinforcement fine-tuning for 12 additional models

This article announces that Amazon SageMaker AI now supports serverless reinforcement fine-tuning for 12 additional open-weight models.

• Newly supported models include Qwen, DeepSeek, Llama, and gpt-oss variants
• Enables fine-tuning without provisioning or managing infrastructure
• Supports supervised fine-tuning (SFT), direct preference optimization (DPO), and reinforcement fine-tuning (RFT)
• RLVR improves accuracy on verifiable tasks like code generation and math
• RLAIF uses AI-generated feedback for quality and safety alignment
• Available in US East, US West, Asia Pacific, and EU regions
• Pay-per-use pricing model with no cluster setup required

SageMaker AI now enables serverless fine-tuning of 12 additional models using advanced reinforcement techniques for domain-specific reasoning tasks.