Updates to Anthropic Console Improve Prompt Quality and Collaboration, Featuring Support for Claude 3.7 Sonnet

Anthropic is announcing that it has redesigned the Anthropic Console—a platform for developers to create and explore prompts for the Anthropic API and Claude—to better serve as a centralized environment for collaboratively building, testing, and iterating on AI deployments. Offering advancements in prompt quality, collaboration, and extended thinking, the new version of Anthropic Console is now available to all users.

Regarding prompt quality, Anthropic Console builds upon its existing capacity to streamline AI development with tools that write, evaluate, and optimize prompts more efficiently. These are broken down into four main components:

• Write better prompts through Workbench, which offers an interactive environment for testing API calls, helping to more effectively structure them and incorporate examples and integrate external tools
• Automatically generate prompts by describing what you want to achieve, enabling Claude to create an effective, precise, reliable prompt through various prompt engineering techniques, including chain-of-thought reasoning
• Evaluate model responses against real-world scenarios with automatic test case generation and side-by-side output comparison, allowing users to easily run test suites, grade response quality, and make data-driven decisions about which prompts to deploy
• Automatically improve prompt with advanced prompt engineering techniques, ideal for adapting prompts originally written for AI models or those that were hand-written

When prompts are finalized, users receive a production-ready API call that can be immediately deployed.

Collaboration is another key piece of the updated Anthropic Console, aiming to remediate the challenges of cross-domain and -department collaboration for improving prompt results. With shareable prompts, various teams can develop, refine, and standardize prompts within a centralized environment. Anthropic Console also enables a shared library of prompts from which team members can access and collaborate, making the process of establishing consistency and best practices across Claude-powered apps more intuitive, according to Anthropic.

Anthropic Console now also features support for Claude 3.7 Sonnet, which Anthropic describes at its latest and most intelligent model to date. With support for Claude 3.7 Sonnet—which can produce near-instant responses or extended, step-by-step thinking that is transparent to the user—Anthropic Console allows for prompts to be optimized for this “extended thinking” model. Anthropic Console can also control Claude 3.7 Sonnet’s “budget” for thinking, determining how long the model takes to deliver a response.

To learn more about Anthropic Console, please visit https://www.anthropic.com/.