---
title: "DeepSeekMath-V2 — paper by Shao et al. (2025)"
description: "Reinforcement learning that only checks final answers has let LLMs master answer-based contests… 20 references, cited by 32 papers here."
authors: ["Zhihong Shao", "Yuxiang Luo", "Chengda Lu", "Z.Z. Ren", "Jiewen Hu", "Tian Ye", "Zhibin Gou", "Shirong Ma", "Xiaokang Zhang"]
year: 2025
venue: "arXiv:2511.22570"
source: "https://arxiv.org/abs/2511.22570"
doi: "10.48550/arxiv.2511.22570"
arxiv: "2511.22570"
licence: "arXiv — non-exclusive licence to distribute"
licence_url: "http://arxiv.org/licenses/nonexclusive-distrib/1.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/shao-2025-deepseekmath-v2-self-verifiable/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/shao-2025-deepseekmath-v2-self-verifiable/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2025](https://md.research.elamur.ai/free/en/papers/2025/page/3/) / DeepSeekMath-V2: Towards Self-Verifiable Mathematical Reasoning

# DeepSeekMath-V2: Towards Self-Verifiable Mathematical Reasoning

**Authors:** [Zhihong Shao](https://md.research.elamur.ai/free/en/authors/zhihong-shao/), Yuxiang Luo, [Chengda Lu](https://md.research.elamur.ai/free/en/authors/chengda-lu/), [Z.Z. Ren](https://md.research.elamur.ai/free/en/authors/z-z-ren/), Jiewen Hu, [Tian Ye](https://md.research.elamur.ai/free/en/authors/tian-ye/), [Zhibin Gou](https://md.research.elamur.ai/free/en/authors/zhibin-gou/), [Shirong Ma](https://md.research.elamur.ai/free/en/authors/shirong-ma/) and [Xiaokang Zhang](https://md.research.elamur.ai/free/en/authors/xiaokang-zhang/) · **Year:** 2025 · **Venue:** arXiv:2511.22570 · **Licence:** [arXiv — non-exclusive licence to distribute](http://arxiv.org/licenses/nonexclusive-distrib/1.0/)

**Links:** [Publisher's page](https://arxiv.org/abs/2511.22570) · [PDF](https://arxiv.org/pdf/2511.22570) · [DOI 10.48550/arxiv.2511.22570](https://doi.org/10.48550/arxiv.2511.22570) · [arXiv 2511.22570](https://arxiv.org/abs/2511.22570) · [OpenAlex](https://openalex.org/W4416947242)

**Citations:** 1 citations (OpenAlex, 2026-10-14) · [32](https://md.research.elamur.ai/free/en/papers/shao-2025-deepseekmath-v2-self-verifiable/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
Reinforcement learning that only checks final answers has let LLMs master answer-based contests like AIME and HMMT. However, it cannot tell sound reasoning from lucky guesses and gives no training signal for theorem proving. DeepSeekMath-V2, built on DeepSeek-V3.2-Exp-Base, addresses this by teaching a model to check proofs written in natural language. The authors first train a verifier with RL on expert-graded proofs for olympiad problems collected from AoPS. They then add a meta-verifier that penalizes invented flaws in the verifier's critiques. The verifier acts as a reward model for a proof generator, which must also grade its own output honestly. This pushes the generator to find and repair errors before answering. As the generator improves, extra verification compute auto-labels difficult new proofs, so the verifier can be retrained without more human annotation. With scaled test-time search, the single resulting model reaches gold-medal level on IMO 2025 and CMO 2024 and scores 118/120 on Putnam 2024. The work suggests that self-checking LLMs are a viable path toward reliable mathematical AI, including on problems with no reference solutions.

## Key points
- **Training data:** 17,503 proof-oriented problems were crawled from Art of Problem Solving (AoPS) contest collections. Experts scored candidate proofs on a three-level scale (0, 0.5, 1).
- **Meta-verification:** This step raised the average quality of the verifier's analyses from 0.85 to 0.96 on a validation split, while the accuracy of predicted proof scores stayed the same.
- **Generator reward:** The reward weights proof quality at α = 0.76 and the accuracy of the generator's self-assessment at β = 0.24.
- **Competition results:**
  - IMO 2025: 5 of 6 problems solved.
  - CMO 2024: 4 problems solved plus partial credit on another.
  - Putnam 2024: 118/120, above the best human score of 90.
  - Test-time search ran up to 16 refinement rounds with 64 verification passes per candidate proof.
- **IMO-ProofBench:** On the basic set, the model outperforms DeepMind's DeepThink (IMO Gold); on the advanced set, it remains competitive. Model and method are released on GitHub (deepseek-ai/DeepSeek-Math-V2).

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [🧮 Computational Theory and Mathematics](https://md.research.elamur.ai/free/en/shelves/computer-science/computational-theory-and-mathematics/) · 321 works

🧭 Topics (1): [🧮 Reasoning](https://md.research.elamur.ai/free/en/topics/reasoning/) · 1,685 works

## Cite

Cite as: Zhihong Shao, Yuxiang Luo, Chengda Lu, Z.Z. Ren, Jiewen Hu, Tian Ye, Zhibin Gou, Shirong Ma and Xiaokang Zhang. 2025. DeepSeekMath-V2: Towards Self-Verifiable Mathematical Reasoning. arXiv:2511.22570.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@misc{shao2025deepseekmath,
  title = {{DeepSeekMath-V2: Towards Self-Verifiable Mathematical Reasoning}},
  author = {Zhihong Shao, Yuxiang Luo, Chengda Lu, Z.Z. Ren, Jiewen Hu, Tian Ye, Zhibin Gou, Shirong Ma and Xiaokang Zhang},
  year = {2025},
  doi = {10.48550/arxiv.2511.22570},
  eprint = {2511.22570},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2511.22570}
}
```

- [← previous Culture is Not Trivia: Sociocultural Theory for Cultural NLP](https://md.research.elamur.ai/free/en/papers/zhou-2025-culture-is-not-trivia-sociocultural/)
- [next → MiMo-VL Technical Report](https://md.research.elamur.ai/free/en/papers/team-2025-llm-mimo-vl-technical-report/)
- [↑ back to papers of 2025](https://md.research.elamur.ai/free/en/papers/2025/page/3/)

## References

20 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. AlphaProof and A. teams. Ai achieves silver-medal standard solving international mathematical olympiad problems, 2024. URL [https://deepmind.google/blog/ai-solves-imo-pro](https://deepmind.google/blog/ai-solves-imo-pro) blems-at-silver-medal-level . [↗](https://deepmind.google/blog/ai-solves-imo-pro)
2. L. Chen, J. Gu, L. Huang, W. Huang, Z. Jiang, A. Jie, X. Jin, X. Jin, C. Li, K. Ma, C. Ren, J. Shen, W. Shi, T. Sun, H. Sun, J. Wang, S. Wang, Z. Wang, C. Wei, S. Wei, Y. Wu, Y. Wu, Y. Xia, H. Xin, F. Yang, H. Ying, H. Yuan, Z. Yuan, T. Zhan, C. Zhang, Y. Zhang, G. Zhang, T. Zhao, J. Zhao, Y. Zhou,… [find](https://scholar.google.com/scholar?q=L%2E+Chen%2C+J%2E+Gu%2C+L%2E+Huang%2C+W%2E+Huang%2C+Z%2E+Jiang%2C+A%2E+Jie%2C+X%2E+Jin%2C+X%2E+Jin%2C+C%2E+Li%2C+K%2E+Ma%2C+C%2E+Ren%2C+J%2E+Shen%2C+W%2E+Shi%2C+T%2E+Sun%2C+H%2E+Sun%2C+J%2E+Wang%2C+S%2E+Wang%2C+Z%2E+Wang%2C+C%2E+Wei%2C+S%2E+Wei%2C+Y%2E+Wu%2C+Y%2E+Wu%2C+Y%2E+Xia%2C+H%2E+Xin%2C+F%2E)
3. Y. Chervonyi, T. H. Trinh, M. Olsák, X. Yang, H. Nguyen, M. Menegali, J. Jung, V. Verma, Q. V. Le, and T. Luong. Gold-medalist performance in solving olympiad geometry with alphageometry2. CoRR, abs/2502.03544, 2025. doi: 10.48550/ARXIV.2502.03544. URL https://doi.org/10.48550/arXiv.2502.03544 . [↗](https://doi.org/10.48550/arXiv.2502.03544)
4. L. M. de Moura, S. Kong, J. Avigad, F. van Doorn, and J. von Raumer. The lean theorem prover (system description). In A. P. Felty and A. Middeldorp, editors, Automated Deduction - CADE-25 - 25th International Conference on Automated Deduction, Berlin, Germany, August 1-7, 2015, Proceedings, volume 9… [find](https://scholar.google.com/scholar?q=L%2E+M%2E+de+Moura%2C+S%2E+Kong%2C+J%2E+Avigad%2C+F%2E+van+Doorn%2C+and+J%2E+von+Raumer%2E+The+lean+theorem+prover+%28system+description%29%2E+In+A%2E+P%2E+Felty+and+A%2E+Middeldorp%2C+editors%2C+Automated+Deduction+%2D+CADE%2D25+%2D+25th+Inter)
5. Springer, 2015. doi: 10.1007/978- 3- 319- 21401- 6\\_26. URL https://doi.org/10.1007/978-3-319-21401-6_26 . — [doi:10.1007/978-](https://doi.org/10.1007/978-) [↗](https://doi.org/10.1007/978-3-319-21401-6_26)
6. G. DeepMind. Gemini 2.5 pro, 2025. URL [https://deepmind.google/models/gemini/pr](https://deepmind.google/models/gemini/pr) o . [↗](https://deepmind.google/models/gemini/pr)
7. DeepSeek-AI. Deepseek-v3.2-exp: Boosting long-context efficiency with deepseek sparse attention, 2025. [find](https://scholar.google.com/scholar?q=DeepSeek%2DAI%2E+Deepseek%2Dv3%2E2%2Dexp%3A+Boosting+long%2Dcontext+efficiency+with+deepseek+sparse+attention%2C+2025%2E)
8. J. Dekoninck, I. Petrov, K. Minchev, M. Balunovic, M. T. Vechev, M. Marinov, M. Drencheva, L. Konova, M. Shumanov, K. Tsvetkov, N. Drenchev, L. Todorov, K. Nikolova, N. Georgiev, V. Kalinkova, and M. Ismoldayev. [The open proof corpus: A large-scale study of llm-generated mathematical proofs](https://md.research.elamur.ai/free/en/papers/dekoninck-2025-open-proof-corpus-large-scale/). CoRR, a…
9. D. Guo, D. Yang, H. Zhang, J. Song, P. Wang, Q. Zhu, R. Xu, R. Zhang, S. Ma, X. Bi, X. Zhang, X. Yu, Y. Wu, Z. F. Wu, Z. Gou, Z. Shao, Z. Li, Z. Gao, A. Liu, B. Xue, B. Wang, B. Wu, B. Feng, C. Lu, C. Zhao, C. Deng, C. Ruan, D. Dai, D. Chen, D. Ji, E. Li, F. Lin, F. Dai, F. Luo, G. Hao, G. Chen, G.… — [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://md.research.elamur.ai/free/en/papers/deepseek-ai-2025-deepseek-r1-incentivizing-reasoning-capability/)
10. Y. Huang and L. F. Yang. Gemini 2.5 pro capable of winning gold at IMO 2025. CoRR, abs/2507.15855, 2025. doi: 10.48550/ARXIV.2507.15855. URL https://doi.org/10 .48550/arXiv.2507.15855 . — [doi:10.48550/arxiv.2507.15855](https://doi.org/10.48550/arxiv.2507.15855) [find](https://scholar.google.com/scholar?q=Y%2E+Huang+and+L%2E+F%2E+Yang%2E+Gemini+2%2E5+pro+capable+of+winning+gold+at+IMO+2025%2E+CoRR%2C+abs%2F2507%2E15855%2C+2025%2E+doi%3A+10%2E48550%2FARXIV%2E2507%2E15855%2E+URL+https%3A%2F%2Fdoi%2Eorg%2F10+%2E48550%2FarXiv%2E2507%2E15855+%2E+%E2%80%94+%5Bdoi%3A10%2E4855)
11. A. Q. Jiang, S. Welleck, J. P. Zhou, T. Lacroix, J. Liu, W. Li, M. Jamnik, G. Lample, and Y. Wu. [Draft, sketch, and prove: Guiding formal theorem provers with informal proofs](https://md.research.elamur.ai/free/en/papers/jiang-2022-draft-sketch-prove-guiding-formal/). In The Eleventh International Conference on Learning Representations, ICLR 2023, Kigali, Rwanda, May 1-5, 2023.
12. OpenReview.net, 2023. URL [https://openreview.net/forum?id=SMa9EAovKMC](https://openreview.net/forum?id=SMa9EAovKMC) . [↗](https://openreview.net/forum?id=SMa9EAovKMC)
13. T. Luong and E. Lockhart. Advanced version of gemini with deep think officially achieves gold-medal standard at the international mathematical olympiad, 2025. URL [https://goo.gle/imo-gold](https://goo.gle/imo-gold) . [↗](https://goo.gle/imo-gold)
14. T. Luong, D. Hwang, H. H. Nguyen, G. Ghiasi, Y. Chervonyi, I. Seo, J. Kim, G. Bingham, J. Lee, S. Mishra, A. Zhai, C. H. Hu, H. Michalewski, J. Kim, J. Ahn, J. Bae, X. Song, T. H. Trinh, Q. V. Le, and J. Jung. [Towards robust mathematical reasoning](https://md.research.elamur.ai/free/en/papers/luong-2025-robust-mathematical-reasoning/). In Proceedings of the 2025 Conference on Empirical…
15. OpenAI. Learning to reason with llms, 2024. URL [https://openai.com/index/learnin](https://openai.com/index/learnin) g-to-reason-with-llms/ . [↗](https://openai.com/index/learnin)
16. OpenAI. Introducing gpt-5, 2025. URL [https://openai.com/index/introducing-gpt-5](https://openai.com/index/introducing-gpt-5) . [↗](https://openai.com/index/introducing-gpt-5)
17. L. C. Paulson. Isabelle - A Generic Theorem Prover (with a contribution by T. Nipkow), volume 828 of Lecture Notes in Computer Science. Springer, 1994. ISBN 3-540-58244-4. doi: 10.1007/BFB0030541. URL https://doi.org/10.1007/BFb0030541 . — [doi:10.1007/bfb0030541](https://doi.org/10.1007/bfb0030541) [↗](https://doi.org/10.1007/BFb0030541)
18. Z. Z. Ren, Z. Shao, J. Song, H. Xin, H. Wang, W. Zhao, L. Zhang, Z. Fu, Q. Zhu, D. Yang, Z. F. Wu, Z. Gou, S. Ma, H. Tang, Y. Liu, W. Gao, D. Guo, and C. Ruan. [Deepseek-prover-v2: Advancing formal mathematical reasoning via reinforcement learning for subgoal decomposition](https://md.research.elamur.ai/free/en/papers/ren-2025-deepseek-prover-v2-advancing-formal/). CoRR, abs/2504.21801, 2025…
19. Z. Shao, P. Wang, Q. Zhu, R. Xu, J. Song, M. Zhang, Y. K. Li, Y. Wu, and D. Guo. [Deepseekmath: Pushing the limits of mathematical reasoning in open language models](https://md.research.elamur.ai/free/en/papers/shao-2024-deepseekmath-pushing-limits-mathematical-reasoning/). CoRR, abs/2402.03300, 2024. doi: 10.48550/ARXIV.2402.03300. URL https://doi.org/10 .48550/arXiv.2402.03300 .
20. T. H. Trinh, Y. Wu, Q. V. Le, H. He, and T. Luong. [Solving olympiad geometry without human demonstrations](https://md.research.elamur.ai/free/en/papers/trinh-2024-solving-olympiad-geometry-without-human/). Nat., 625(7995):476–482, 2024. doi: 10.1038/S41586-023-06747-5. URL https://doi.org/10.1038/s41586-023-06747-5 . [↗](https://doi.org/10.1038/s41586-023-06747-5)

*Updated 2 November 2026*