LLM Speedrunner: Eval for frontier models to reproduce scientific findingsgithub.com2 points·zerojames··0 commentsOpen articleSaveView on HN