Amazon Web Services announced the creation of a marketplace within Amazon Bedrock from which developers can easily find and install popular, emerging, and specialized language models. Although the company’s fully managed service for building and scaling generative AI applications already supports multiple models, this marketplace enables developers to swap out the LLM without requiring direct application modifications.
A key differentiator between what AWS previously had and now is that today’s Amazon Bedrock marketplace is geared toward specialized models rather than the larger models most developers might flock to.
In addition, AWS believes this will lead to better development workflow thanks to the unified console experience within this new marketplace.
Disclosure: I attended Amazon's 2024 re:Invent as a guest, with a portion of my travel expenses covered by the company. However, Amazon had no influence over the content of this post—these thoughts are entirely my own.Some of the models available within the Amazon Bedrock marketplace include Mistral AI’s Mistral NeMo Instruct 2407, Technology Innovation Institute’s Falcon RW 1B, Nvidia’s NIM microservices, Writer’s Palmyra-Fin which is made for the financial industry, Upstage’s Solar Pro for translation, Camb.ai’s text-to-audio MARS6, and EvolutionaryScale’s ESM3, a biology-specific generative model.
This feature launch puts AWS in parity with competitors with similar offerings, such as Hugging Face, OpenAI via OpenRouter, Google’s Cloud and Vertex AI, and Microsoft’s Azure AI Studio.
According to AWS, when developers identify a model they want to use, they select the appropriate infrastructure and deploy it on AWS through fully managed endpoints. Then, they integrate the chosen model with Amazon Bedrock’s unified APIs to access guardrails, agents, and the platform’s security and privacy features.
This is the latest update AWS has made to Amazon Bedrock in as many days and comes during the company’s annual re:Invent developer show. Yesterday, it revealed that it was adding automated reasoning, multi-agent orchestration, and model distillation to the offering.
Featured Image: The Amazon Web Services logo hangs atop the expo hall during the company's 2024 re:Invent conference in Las Vegas, Nevada. Photo credit: Ken Yeung
Leave a Reply
You must be logged in to post a comment.