Recent advancements in LLMs such as OpenAI-o1, DeepSeek-R1, and Kimi-1.5 have significantly improved their performance on complex mathematical reasoning tasks. Reinforcement Learning with Verifiable ...