rest-stress-testing-reasoning

Evaluate large reasoning models under stress by asking multiple problems simultaneously, revealing true multi-tasking capacity and context management. Use to identify robustness gaps invisible in single-problem evaluations and discriminate model capabilities beyond traditional benchmarks.

Independent GitHub snapshot · Automated classification · Original documentation governs