Amazon Web Services has introduced a redesigned console experience for Amazon Bedrock, the company’s managed platform for building with foundation models. The new interface is aimed at developers working with Anthropic- and OpenAI-compatible APIs, giving them a more direct way to test models, compare options, and move from evaluation to production.
AWS said the updated console is built around its next-generation inference engine and supports the latest GPT, Claude, and open-weight models through the OpenAI Responses API, OpenAI Chat Completions API, and Anthropic Messages API. The company is positioning the console as a streamlined workspace for developers who want to experiment with models and then take the same project into deployment.
The redesign centers on three main features. First, AWS added a new model catalog that lets users browse available models and compare them side by side. The catalog surfaces details such as supported capabilities, modality, context window, service quotas, token information, pricing, and regional availability. Users can compare up to three models at once, which AWS says should reduce the need to jump between documentation and separate limit calculators.
Second, the console now uses a project-based workflow. Developers can create a project, assign models, configure API keys, run evaluations, and review usage data from a single dashboard. AWS says the layout reflects the lifecycle of a generative AI application, from early testing through ongoing monitoring.
The project dashboard also includes usage insights such as token consumption, requests per minute, and tokens per request. AWS says those metrics can help developers make decisions about model choice, prompt tuning, and workload consistency.
A third major addition is what AWS calls live documentation. The console can automatically prefill code samples, SDK snippets, and API references with the project’s settings, including model selection, region, endpoint URL, and API key reference. That means developers can copy a snippet into their application and run it without having to manually edit the details first.
AWS also says the new console can help with application setup. In the Getting Started area, users can migrate existing code, build a new application with Anthropic or OpenAI SDKs, or connect an AI coding assistant to Bedrock. The console offers environment code for terminal testing, as well as instructions for saving configuration in a .env file.
For developers using AI coding tools, the console includes guidance for connecting agents such as Claude Code, Cline, Codex, Cursor, and OpenCode to the Bedrock mantle endpoint. The instructions cover installation, authentication through AWS IAM credentials or a Bedrock API key, and environment variable setup.
AWS says the new experience is separate from the existing Bedrock console, which remains available for fully managed features such as Agents, Knowledge Bases, Guardrails, fine-tuning, and the InvokeModel and Converse APIs that run on the bedrock-runtime endpoint.
The company said the new console is available in all AWS Regions where the bedrock-mantle endpoint is offered, including several locations across North America, Europe, Asia Pacific, and South America. AWS is inviting users to try the interface and share feedback through re:Post or standard support channels.
The launch reflects AWS’s effort to simplify how developers access competing model families through a single cloud platform. By combining model discovery, project management, and code-ready documentation in one place, the company is aiming to make Bedrock more approachable for teams that are comparing AI providers while building production applications.