{"slug": "mini-lean-agent-a-coding-agent-baseline-for-lean", "title": "mini-Lean-Agent: A Coding-Agent Baseline for Lean", "summary": "A paper submitted to arXiv on 12 October 2026 introduces mini-Lean-agent, a deliberately minimal Lean coding agent whose only model-visible tool is bash, and reports that it solves 12/12 Putnam 2025 problems with Astra as the base model at an average cost of $1.28 per problem. With Gemini 3.8 Flash as the base model, mini-Lean-agent solves 7/12 problems under a $2-per-problem limit, versus 4/12 for AxProverBase, which the authors compare against alongside Numina-Lean-Agent. The authors argue the results show a generic coding agent with a minimal harness can already achieve strong performance on Lean benchmarks, and propose mini-Lean-agent as a standard baseline for evaluating theorem-proving infrastructure.", "body_md": "# Computer Science > Artificial Intelligence\n\n  [Submitted on 12 Oct 2026]\n\n# Title:Do Theorem Provers Need Theorem-Proving Agents? mini-Lean-agent: A Coding-Agent Baseline for Lean\n\n[View PDF](https://arxiv.org/pdf/2610.17331)\n\n[HTML (experimental)](https://arxiv.org/html/2610.17331v1)\n\nAbstract:General-purpose coding agents use standard shell and file interactions, yet formal theorem-proving agents often rely on specialized, ad-hoc tooling. This makes it difficult to determine which components of theorem-proving agents provide benchmark gains: an improved base model, a specialized harness, a specific tool, or a combination of any of them. We introduce mini-Lean-agent, a deliberately minimal agent for Lean tasks whose only model-visible tool is bash. The model must use bash to inspect and edit the repository, search local libraries, invoke Lean, and manage its own intermediate work, while a separate trusted evaluator verifies the final artifact. We compare mini-Lean-agent against AxProverBase and Numina-Lean-Agent on the Putnam 2025 benchmark. Experiments with Gemini 3.8 Flash as the base model show that mini-Lean-agent solves 7/12 problems at \\$2 per problem limit, compared with AxProverBase's 4/12. With Astra as the base model, mini-Lean-agent solves 12/12 problems at an average cost of \\$1.28 per problem. This suggests that a generic coding agent with a minimal harness can already achieve strong performance on Lean benchmarks. Our results motivate using mini-Lean-agent as a standard baseline when evaluating theorem-proving infrastructure.\n    \n\n### Additional Features\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer \n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers \n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps \n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations \n\n*(*[What are Smart Citations?](https://www.scite.ai/))\n# Code, Data and Media Associated with this Article\n\nalphaXiv \n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers \n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub \n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub \n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face \n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast \n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))\n# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower \n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender \n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))\n# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [**Learn more about arXivLabs**](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/mini-lean-agent-a-coding-agent-baseline-for-lean", "canonical_source": "https://arxiv.org/abs/2610.17331", "published_at": "2026-10-21 06:33:13+00:00", "updated_at": "2026-10-21 07:11:23.430794+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-agents", "ai-research", "large-language-models", "ai-tools"], "entities": ["mini-Lean-agent", "Lean", "AxProverBase", "Numina-Lean-Agent", "Gemini 3.8 Flash", "Astra", "arXiv", "Putnam 2025"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/mini-lean-agent-a-coding-agent-baseline-for-lean", "markdown": "https://wpnews.pro/news/mini-lean-agent-a-coding-agent-baseline-for-lean.md", "text": "https://wpnews.pro/news/mini-lean-agent-a-coding-agent-baseline-for-lean.txt", "jsonld": "https://wpnews.pro/news/mini-lean-agent-a-coding-agent-baseline-for-lean.jsonld"}}