[DeepSeek launches DeepSeekMath-V2 model] According to Zhitong Finance, DeepSeek has launched a new mathematical reasoning model, DeepSeekMath-V2, which adopts a self-verifiable training framework. The model is built on DeepSeek-V3.2-Exp-Base and automatically reviews generated mathematical proofs through an LLM validator, continuously optimizing performance using high-difficulty samples.