Bring more intelligence to everyday work with GPT-6 Sol and GPT-6 Luna on Amazon Bedrock

Amazon has made GPT-6 Sol and GPT-6 Luna, two powerful AI models, available on its Amazon Bedrock service. These models are designed to enhance the intelligence and efficiency of workloads, particularly those that require complex reasoning or frequent use.

GPT-6 Sol is particularly useful for recurring tasks in software development and operations. It can perform tasks such as writing code, debugging, refactoring, and analyzing data. The model is noted for making fewer factual mistakes than its predecessor, GPT-5.6 Sol, and provides clearer communication about its work and results. This makes it practical for use throughout the development cycle.

On the other hand, GPT-6 Luna is optimized for handling large volumes of work. It can be used for tasks like extracting information from large document collections, summarizing material, classifying inputs, and answering questions. Efficiency at volume is achieved through consistent outputs and adjustable reasoning effort per request. GPT-6 Luna also supports prompt caching, allowing it to reuse context across requests, thus improving efficiency.

Both models run on a high-performance inference engine built for security and reliability at scale. Amazon Bedrock provides a foundation for these models with features like AWS Identity and Access Management (IAM) policies for model access, audit logging through AWS CloudTrail, and VPC endpoints to keep traffic within network boundaries. The inference runs on hardware-isolated infrastructure with no operator access, ensuring data security.

To use GPT-6 Sol and GPT-6 Luna, you can access them through the Amazon Bedrock console or programmatically using supported Amazon Bedrock APIs. This allows organizations to bring advanced AI capabilities into their production systems with the appropriate level of performance, control, and flexibility.