Amazon Web Services (AWS) announced that it will offer the DeepSeek-R1 model through its Amazon Bedrock and Amazon SageMaker AI platforms.
DeepSeek unveiled the DeepSeek-V3 model in December 2024, followed by the release of DeepSeek-R1, DeepSeek-R1-Zero, and DeepSeek-R1-Distill models on January 20. On January 27, they introduced the vision-based Janus-Pro 7B model.
AWS is supporting its customers in leveraging the DeepSeek-R1 model through various approaches.
An undisclosed representative of AWS explained that Amazon Bedrock enables seamless integration of pre-trained models via API (Application Programming Interface), while Amazon SageMaker AI facilitates customized training and distribution.
It is also recommended that Amazon Bedrock guardrails be utilized to enhance security and scalability. This feature allows for monitoring generative AI applications’ input and output data and filtering out harmful content.
The DeepSearch-R1 model can be deployed using the AWS platform in four ways: through the Amazon Bedrock Marketplace, Amazon SageMaker Jumpstart, Amazon Bedrock Custom Model Import, and AWS EC2 Trn1 instances.