DeepSeek launches DeepSeekMath-V2 model

Jinse Finance reported that DeepSeek has launched a new mathematical reasoning model, DeepSeekMath-V2, which adopts a self-verifiable training framework. This model is built on DeepSeek-V3.2-Exp-Base, automatically reviews generated mathematical proofs through an LLM verifier, and continuously optimizes performance using high-difficulty samples.