Anthropic published a guest post by Prof. Matthew Schwartz describing BootLoops, an open-source toolkit that acts as a harness for Claude on quantitative science problems. Schwartz reports that Claude Opus 4.5 performed like a strong graduate student at 20 times the speed, but required heavy correction. By targeting "Claude-shaped" problems suited to the model's strengths, BootLoops found technically correct connections across ecology, population genetics, and other fields, which Schwartz refined with domain experts.
No score is assigned. Sources and their independence are shown in the citation chain below.