Search
On the evening of November 27, DeepSeek made a significant stride by introducing its latest mathematical reasoning model, DeepSeekMath-V2. Leveraging a self-verifiable training framework, this model is constructed on the foundation of DeepSeek-V3.2-Exp-Base. It harnesses the power of a large language model (LLM) verifier to automatically scrutinize generated mathematical proofs. By utilizing challenging samples, the model undergoes a process of continual performance refinement. Notably, the model attained gold medal-standard excellence in both the 2025 International Mathematical Olympiad (IMO 2025) and the 2024 China Mathematical Olympiad (CMO 2024). Moreover, it excelled in the 2024 Putnam Mathematical Competition (Putnam 2024), securing an impressive score of 118 points out of a possible 120.
-
C114 Communication Network -
Communication Home
