S1-Bench: A Simple Benchmark for Evaluating System 1 Thinking Capability of Large Reasoning Models Paper • 2504.10368 • Published Apr 14 • 21