Show HN: Do Codex skills save tokens? A six-run task-size benchmark
Six controlled GPT-5.6-sol runs reveal a task-size boundary worth replicating.
Summary not yet generated for this article.
Comments (1 of ) ⌛