OpenAI Says GPT-6 Models Are Better at COBOL, Fortran and Other Low-Resource Languages

OpenAI published a blog post dated October 19, 2026, on coding in low-resource programming languages such as COBOL, Fortran and Visual Basic. The company says these languages have less public code and documentation for models to learn from than languages like Python and JavaScript, yet still run critical systems. An OpenAI post on X about the same topic went out at 06:30 Beijing time on October 20.

OpenAI says it measured progress with two of its own evaluations. MultiLangBench asks a model to build a SQLite-compatible database engine from scratch in each of 20 programming languages. A separate COBOL migration evaluation asks a model to migrate COBOL programs to C++, Java, Kotlin, Python or Rust while preserving behavior.

MultiLangBench

According to OpenAI's own figures, at maximum reasoning effort GPT-6 Astra reaches an average test pass rate of 87.4% across the 20 languages, up from 26.8% for GPT-5.6 Sol, with gains in every language tested. Examples OpenAI lists include COBOL rising from 10.9% to 72.1%, OCaml from 25.6% to 85.8%, Fortran from 21.4% to 83.5%, and Visual Basic from 27.5% to 89.4%. OpenAI says Astra scores above 83% in 19 of the 20 languages, with COBOL the hardest at 72%.

GPT-6 Luna, OpenAI says, also beats GPT-5.6 Sol on every language in the same suite, reaching a 40.3% average pass rate at maximum reasoning effort, at an estimated average API cost about 78% lower ($0.77 per sample versus $3.48 for GPT-5.6 Sol).

OpenAI has released the MultiLangBench evaluation setup, test suite and per-language results on GitHub so developers can inspect the tests and the scores.

COBOL migration

On OpenAI's COBOL migration evaluation, GPT-6.1 Sol reaches 86.6% accuracy at low reasoning effort, compared with 5.4% for GPT-5.6 Sol at the same setting, and OpenAI says that run costs about 39% less. At higher reasoning effort, OpenAI reports 97.9% accuracy, which it describes as nearly saturating the evaluation. OpenAI notes that applying the approach to production software still needs expert review and testing against each system's business rules.