Effort as Ceiling, Not Dial: Reasoning Budget Does Not Modulate Cognitive Cost Alignment Between Humans and Large Reasoning Models

In three linesLarge Reasoning Models generate traces aligned with human reaction times, but this alignment persists regardless of inference-time reasoning budget. Study across GPT-OSS-20B and GPT-OSS-120B: three effort levels, six cognitive tasks. Token allocation tracks fine-grained human difficulty patterns and reflects a structure crystallized at training time, not modulated in real-time.

Summary generated by Claude — human-verified