Amazon Web Services (AWS) recently announced that the DeepSeek-R1 large language model is now available as a fully managed, serverless service on Amazon Bedrock. As the first cloud provider to offer this model as a fully managed service, AWS further expands how customers can utilize DeepSeek-R1 and its distilled versions.
With the fully managed service approach, customers can easily deploy DeepSeek-R1 in enterprise settings without dealing with complex technical setups or maintenance. Customers can leverage the model to solve complex problems, write code, and process and analyze data. Simultaneously, customers benefit from AWS enterprise-grade security safeguards, including data encryption, strict access control, and Amazon Bedrock Guardrails to detect and mitigate model hallucinations.
It's noteworthy that DeepSeek-R1 was previously available on the Amazon Bedrock Marketplace, where customers could run the model using their own managed infrastructure. Furthermore, customers can upload their fine-tuned distilled DeepSeek-R1 Llama models via Amazon Bedrock's custom model import feature and run them as fully managed models. Since its launch in late January of this year, thousands of customers have deployed DeepSeek-R1 models using this feature.