Anthropic has released dynamic workflows in Claude Code, a new feature that lets the AI break down large coding tasks into hundreds of smaller subtasks and run them in parallel across multiple agents. The system checks results before presenting a final answer, designed for work that is too complex or large for a single AI pass, including codebase-wide bug investigations, security reviews, and large migrations.

What Are Dynamic Workflows and How Do They Work?

Dynamic workflows represent a shift in how AI coding assistants approach complex software engineering problems. When a workflow starts, Claude creates a plan from the user's request and divides the job into smaller pieces. It then assigns those tasks to parallel agents, while separate agents check or challenge the findings before the results are combined. The system is designed for long-running work that can continue for hours or days, with progress saved during the run so interrupted jobs can resume rather than restart.

The feature is available in the Claude Code command-line interface, desktop application, and Visual Studio Code extension for users on Max, Team, and some Enterprise plans. Anthropic is also making the feature available through the Claude API and on Amazon Bedrock, Vertex AI, and Microsoft Foundry.

How to Use Dynamic Workflows for Your Coding Projects

• Enable the Feature: For Max and Team users, dynamic workflows are on by default. Enterprise customers need an administrator to enable the feature through managed settings.
• Start with Limited Tasks: Anthropic recommends starting with a limited task, noting that dynamic workflows can use substantially more tokens than a standard Claude Code session, which may increase costs for users operating under commercial plans.
• Request Confirmation: The first time a workflow is triggered, Claude Code shows the user what is about to run and asks for confirmation before proceeding.
• Use the Ultracode Setting: A new Claude Code setting called ultracode, available through the effort menu, sets the effort level to xhigh and allows Claude to decide automatically when to use a workflow.

Real-World Example: The Bun JavaScript Runtime Migration

One of the most striking examples of dynamic workflows in action comes from the Bun JavaScript runtime project. Jarred Sumner used dynamic workflows to help port Bun from Zig to Rust, producing approximately 750,000 lines of Rust code with 99.8% of the existing test suite passing, completing the work in 11 days from first commit to merge. The workflow operated in multiple stages: one workflow identified Rust lifetimes for struct fields in the Zig codebase, while another wrote Rust files as matching ports of the Zig originals. A further loop drove the build and test process until both ran clean, and a later overnight workflow opened pull requests aimed at reducing unnecessary data copies.

This example points to the kind of large-scale software maintenance and migration work AI developers are increasingly targeting as they try to move beyond coding assistance into more autonomous engineering tasks. However, it is important to note that this work has not yet been put into production.

Why Anthropic Is Betting on Parallel Agent Orchestration

The release reflects broader competition among AI model developers to offer tools that can manage more of the software development cycle, from analysis and planning to implementation and review. Anthropic is placing the feature not only in its own coding products but also across cloud marketplaces run by major technology companies, widening the routes through which customers can use it.

Anthropic framed the feature as a way to tackle tasks that are too large or too complex for a single-agent run, especially in legacy systems where code is spread across large repositories and changes need independent verification. The system can keep iterating until answers converge, rather than relying on a single pass. Anthropic said dynamic workflows can handle "critical work you need checked twice" by using independent attempts at a problem and adversarial agents that try to break the result before the user sees it.

Anthropic

The Broader Context: Enterprise Adoption of Coding Agents

The launch of dynamic workflows comes as both Anthropic and OpenAI are seeing significant enterprise adoption of their coding agent products. According to industry observers, companies are spending $200 or more per month per user on these tools, with some power users accumulating API costs exceeding $1,000 per month per vendor. This represents a meaningful shift in how enterprises are budgeting for AI, with coding agents becoming daily drivers for well-compensated professionals like software engineers.

The token consumption required by dynamic workflows is higher than standard Claude Code sessions, which means users operating under commercial plans or through APIs will see increased costs. However, for large-scale tasks that would otherwise require weeks of manual engineering work, the trade-off may be worthwhile. Anthropic is also releasing Claude Opus 4.8, its newest flagship model, with improved performance on coding benchmarks and pricing that remains unchanged from the previous version.

Dynamic workflows represent a significant step toward autonomous software engineering, though they remain in research preview. As more organizations test the feature on production codebases, the practical constraints and benefits of parallel agent orchestration will become clearer.