OpenAI has announced that its newest model, GPT-6, demonstrates substantial advancements in processing legacy programming languages such as COBOL and Fortran. The model’s performance is highlighted by an impressive 87.4 percent test pass rate across 20 languages, a significant increase from the previous model, GPT-5.6 Terra, which achieved only about 38 percent.
Particularly noteworthy is GPT-6.1 Sol’s performance in COBOL migration, reaching an accuracy of 97.9 percent. This improvement is expected to have meaningful implications for the modernization of critical financial systems that rely on older programming languages, potentially streamlining the migration process and reducing the associated costs.
This leap in performance not only enhances the utility of AI in legacy systems but also opens new avenues for developers and organizations looking to upgrade their technology stacks. As AI continues to evolve, its role in bridging the gap between old and new programming paradigms becomes increasingly vital.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
