AWS Introduces Mistral Large Model to Amazon Bedrock for Enhanced AI Capabilities

Amazon Web Services (AWS) is expanding its Amazon Bedrock platform, a managed service for generative AI development, by adding new model offerings. The latest addition is Mistral Large, now available for developers to enhance their AI applications. This follows the recent inclusion of Mistral 7B and Mixtral 8x7B models in Bedrock.

Vasi Philomin, AWS's Vice President of Generative AI, stated, “By integrating Mistral AI models into Amazon Bedrock, customers gain access to advanced generative AI technologies along with enterprise-grade tools, all within a secure environment.”

Mistral Large, released in February, excels in complex multilingual tasks such as text understanding, transformation, and code generation. It is ranked among the top models available via API and offers native fluency in English, French, Spanish, German, and Italian. With a 32K token context window and precise instruction-following capabilities, it allows developers to establish custom moderation policies and is inherently capable of function calling.

Arthur Mensch, CEO of Mistral AI, emphasized the collaboration with AWS: “Our mission is to make frontier AI ubiquitous. By partnering with the world's leading cloud provider, we can deliver tailored AI solutions to developers globally.”

All Mistral models are now accessible in the U.S. East (N. Virginia), U.S. West (Oregon), and EU-West-3 (Paris) regions. They join models from other providers, including Anthropic, AI21 Labs, Cohere, Meta, Stability AI, and Amazon.

Looking ahead, AWS and Mistral plan to leverage Amazon’s Tranium and Inferentia chips for future foundational model development.

Additionally, Amazon Bedrock is now operational in France, allowing developers to utilize all supported large language models (LLMs) and foundational models while ensuring data security within French borders.