Kimi's new algebraic reasoning model, k0-math, is now online!

On November 25, Kimi's new algebraic reasoning model k0-math was officially launched. Officially, k0-math is the first reasoning ability enhancement model launched by Kimi, adopting new reinforcement learning and chain-of-mind reasoning technology, which greatly improves the ability to solve mathematical puzzles by simulating the thinking and reflective process of the human brain, and can help users complete more challenging mathematical tasks . In addition, k0-math's performance in several math benchmarks is comparable to that of the two publicly available models in the OpenAI o1 series: o1-mini and o1-preview. k0-math's first-generation model outperforms the o1-mini and o1-preview models in four math benchmarks: the middle school exam, the university entrance exam, the graduate school exam, and the MATH, which includes introductory competition questions. preview model.