Large Reasoning Models Fail to Follow Instructions During Reasoning: A Benchmark Study

ReasonIF finds frontier LRMs fail to follow reasoning instructions >75% of the time; introduces a benchmark across languages, formatting, and length.

Raccontata datogether.ai