ScarfBench: Benchmarking AI Agents for Enterprise Java Framework Migration

Modernizing enterprise applications is one of the largest and most expensive software engineering activities organizations undertake. Teams migrate applications across frameworks to improve maintainability, cloud readiness, developer productivity, and access to modern capabilities.

Recent advances in coding agents have sparked excitement around AI-assisted modernization. But an important question remains:

Can AI agents reliably modernize real-world enterprise applications?

Existing software engineering benchmarks have demonstrated impressive progress in bug fixing and code generation, but framework migration presents a fundamentally different challenge. Success requires not only translating code, but also preserving behavior, adapting build systems, and navigating runtime dependencies.

To address this gap, we introduce ScarfBench (Self-Contained Application Refactoring Benchmark) , an open benchmark for evaluating AI agents on cross-framework migration tasks in Enterprise Java.

Unlike traditional benchmarks that compare generated code against reference implementations, ScarfBench evaluates whether migrated applications actually build, deploy, and preserve behavior.