Event ·Multi-source

OpenAI 发布新模型，称其在 COBOL 和 Fortran 上的表现提升

OpenAI 表示新模型在 COBOL 和 Fortran 这类老牌编程语言上的表现有所改善。同时，OpenAI 在 GitHub 上开源了 MultiLangBench 基准，用来测试模型能否用 20 种编程语言构建一个 SQLite 兼容的数据库引擎。这套基准覆盖 COBOL、Fortran 等常见基准较少涉及的语言，为评估老代码维护能力提供了工具。

Current read

OpenAI 的新模型专门照顾了 COBOL、Fortran 这种老语言，还开源了跨 20 种语言的 MultiLangBench，搞遗留系统维护的可以看看。

11 sources63° AI heatLast updated 2026/10/20 01:48:07

Evidence

11 sources

Primary sourceTech in Asia

2026/10/20 01:48:07View source

Related source 1IT之家

2026/10/20 23:45:48View source

Related source 2Decoder

2026/10/20 16:40:29View source

Related source 3OpenAI

2026/10/19 22:30:05View source

Related source 4Rest of World

2026/10/18 13:52:00View source

Related source 5Teortaxes

2026/10/19 06:02:50View source

Related source 6Nikkei Asia

2026/10/20 00:55:53View source

Related source 7量子位

2026/10/20 01:09:43View source

Related source 8arXiv: OpenAI

2026/10/20 09:52:31View source

Related source 9Liputan6 Tekno

2026/10/21 04:11:49View source

Related source 10Artificial Analysis

2026/10/21 15:53:56View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story