What we know

OpenAI says its models now have improved abilities to translate COBOL code. This development aims to assist banks and other organizations in migrating from legacy COBOL systems to modern programming languages. While OpenAI highlights these enhanced capabilities, there is no independent verification yet of how well the models perform on large-scale COBOL translation projects or their adoption in real-world banking environments. OpenAI’s models are positioned to support migration away from legacy COBOL systems.

Why it matters

COBOL remains a foundational language for many banking and financial systems, despite being decades old. Transitioning from COBOL to newer languages is often costly, complex, and prone to errors, creating operational risks. OpenAI’s improved translation models could potentially reduce manual effort and speed up modernization efforts.

What is still unknown

There are no independent tests confirming the accuracy or reliability of OpenAI’s COBOL translation capabilities. Details on supported languages beyond COBOL and timelines for broader availability have not been disclosed. No information is available on how these models handle complex or domain-specific COBOL code, nor on integration with existing migration workflows.