Amazon Bedrock Generative AI Model Offerings at AWS Get Llama 2 and General Availability

After a slow start, AWS has quickly closed the foundation model offering gap

Amazon Bedrock was announced in April, which included limited access to generative AI foundation models from AI21 Labs, Amazon (Titan), Anthropic, and Stability AI. The first three were for large language models (LLM) and the Stability for AI image generation. The solution this week graduated to general availability and added Meta’s Llama 2 open-source LLM. Cohere was also added to the LLM list in recent months.

The company’s generative AI strategy is primarily manifested through AWS, where the Bedrock service is accessed. It includes AI-optimized ASICs that compete with NVIDIA GPUs, applications, and a foundation model (FM) marketplace. A blog post announcing the updates commented:

Amazon Bedrock’s comprehensive capabilities help you experiment with a variety of top FMs, customize them privately with your data using techniques such as fine-tuning and retrieval-augmented generation (RAG), and create managed agents that perform complex business tasks—all without writing any code.

…

Amazon Bedrock offers chat, text, and image model playgrounds. In the chat playground, you can experiment with various FMs using a conversational chat interface. The following example uses Anthropic’s Claude model:

Since Amazon Bedrock is serverless, you don’t have to manage any infrastructure, and you can securely integrate and deploy generative AI capabilities into your applications using the AWS services you are already familiar with.

The Big Bet on Anthropic

Amazon’s Bedrock strategy is to provide the “everything store” for foundation models. That is everything except for OpenAI’s and Google’s offerings. Amazon would likely offer OpenAI models if they are made available beyond Azure. However, that does not appear likely in the near term, so AWS is looking to offer everything else.

The situation also propelled the company into investing up to $4 billion in Anthropic, an emerging OpenAI rival, founded by former employees. You can think of Bedrock as the AWS equivalent of Google Cloud’s Vertex AI model garden. Bringing Anthropic closer to AWS was a move to compete more directly with OpenAI and Google’s PaLM and Gemini offerings.

The Llama is Coming

Adding Meta’s Llama 2 is an important step for AWS, as it is already available through Azure and Google Cloud and has quickly become the most talked about open-source LLM. This will help ensure AWS customers don’t have to go to another cloud to get access to the model. Currently, Amazon lists the Llama 2 13B and 70B parameter models as coming soon.

Governance and Audit Features

As more enterprises move generative AI solutions into production, the importance of governance solutions and processes will rise. AWS is smartly promoting its CloudWatch and CloudTrail offerings to support these needs.

Amazon Bedrock is integrated with Amazon CloudWatch and AWS CloudTrail to support your monitoring and governance needs. You can use CloudWatch to track usage metrics and build customized dashboards for audit purposes. With CloudTrail, you can monitor API activity and troubleshoot issues as you integrate other systems into your generative AI applications. Amazon Bedrock also allows you to build applications that are in compliance with the GDPR and you can use Amazon Bedrock to run sensitive workloads regulated under the U.S. Health Insurance Portability and Accountability Act (HIPAA).