Native Parallel Reasoner: Reasoning in Parallelism via Self-Distilled Reinforcement Learning Paper โข 2512.07461 โข Published Dec 8, 2025 โข 75 โข 4