OpenAI says new models improve COBOL, Fortran work
Culture Index
Score Breakdown
Relevance
3/25
Freshness
25/25
Authority
18/20
Brand Signal
10/15
Depth
2/15
5-Axis Cultural Radar
It is also releasing MultiLangBench on GitHub, an evaluation that tests whether models can build a SQLite-compatible database engine across 20 programming languages.
