Can Open-Weight LLMs Produce Kernel-Verified Coq Proofs? A Pilot Study

• URL: http://arxiv.org/abs/2608.05420v1
• Date: Wed, 05 Aug 2026 21:29:12 GMT
• Title: Can Open-Weight LLMs Produce Kernel-Verified Coq Proofs? A Pilot Study
• Abstract summary: Large language models (LLMs) can generate text that resembles a mathematical proof, but resemblance does not establish correctness.<n>Coq bases its rules on the Calculus of Inductive Constructions, a logical framework that defines which proof steps the system may accept.<n>This pilot study evaluated six open-weight LLMs on the same 100 theorems from CoqStoq, a benchmark derived from real Coq projects.
• Score: 0.20415910628419062
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models (LLMs) can generate text that resembles a mathematical proof, but resemblance does not establish correctness. A formal proof checker verifies whether each proof step follows established logical rules. Coq bases its rules on the Calculus of Inductive Constructions, a logical framework that defines which proof steps the system may accept. This pilot study evaluated six open-weight LLMs on the same 100 theorems from CoqStoq, a benchmark derived from real Coq projects. Each LLM received one attempt per theorem with the temperature set to 0, and Coq checked every proposed proof in the theorem's original project environment. We counted a proof as successful only if the Coq kernel accepted it. Gemma 4 verified 12 of 100 theorems, Llama 3.3 verified 8, and DeepSeek Coder V2 Lite verified 1. Qwen 3.5, Mistral Small 3.1, and GPT-OSS verified none. The 21 successful model-theorem results covered 15 distinct theorems, 11 of which were not solved by a baseline of standard Coq tactics. All verified theorems had short or medium human-written reference proofs; no model verified a theorem with a long reference proof. Because the proof-length analysis was exploratory, this pattern does not establish that proof length caused the difference. For the three models with at least one success, the total generation cost per verified proof ranged from 741 to 36,193 output tokens, 14.9 to 178.0 seconds, and 0.0167 to 0.2000 aggregate GPU hours. We could not calculate these ratios for models with no verified proofs. Across 600 attempts, the models produced 21 kernel-verified proofs, giving an overall success rate of 3.5%. The study reports descriptive differences among the models but does not statistically test whether one model outperforms another. Therefore, the results do not establish a universal ranking of the six models.

Related papers

• Harnessing Code Agents for Automatic Software Verification [0.47154736035631406]
  Large language models (LLMs) promise to generate proofs automatically, yet existing approaches wire a fixed, human-designed proof strategy into the system.<n>We show that imposing such a strategy is unnecessary and limiting.<n>Handing the whole lemma to a general code agent, free to choose its own approach, and wrapping it in a verification harness is both simpler and more effective.
  arXiv  Detail & Related papers  (2026-07-07T14:39:59Z)
• MerLean-Prover: A Recursive Looping Harness for Lean 4 Theorem Proving [0.0]
  MerLean-Prover is an end-to-end Lean4 theorem prover that replaces sorry declarations with kernel-checkable proofs.<n>On FormalQualBench, a benchmark of 23 PhD-qualifying-exam theorems, MerLean-Prover solves 10/23.<n>The harness also transfers to smaller models: Sonnet closes all four tested FormalQualBench problems, and Haiku closes the two short ones.
  arXiv  Detail & Related papers  (2026-05-26T12:49:49Z)
• LiveMathematicianBench: A Live Benchmark for Mathematician-Level Reasoning with Proof Sketches [61.30693283718321]
  We present LiveMathematicianBench, a dynamic multiple-choice benchmark for research-level mathematical reasoning.<n>By grounding evaluation in newly published theorems, it provides a realistic testbed beyond memorized patterns.<n>It employs a proof-sketchguided distractor pipeline that uses high-level proof strategies to construct plausible but invalid answer choices.
  arXiv  Detail & Related papers  (2026-04-02T08:22:17Z)
• Mathematics with large language models as provers and verifiers [1.1029146548022293]
  ChatGPT solves five out of six IMO problems, and close about a third of the sixty-six number theory conjectures in [Cohen, Journal of Sequences, 2025].
  arXiv  Detail & Related papers  (2025-10-11T20:35:25Z)
• Goedel-Prover: A Frontier Model for Open-Source Automated Theorem Proving [72.8626512877667]
  We introduce Goedel-Prover, an open-source language model that achieves state-of-the-art (as of April 5 2025) performance in automated formal proof generation for mathematical problems.<n>First, we train LLMs to convert natural language math problems from the Numina dataset to equivalent formal statements in Lean 4.<n>Next, we develop a large dataset of formal proofs by training a series of provers.<n>Finally, we obtain the dataset Goedel-Pset-v1-solved, which contains proofs for over 800K statements from Goedel-Pset-v1
  arXiv  Detail & Related papers  (2025-02-11T15:27:35Z)
• Formal Theorem Proving by Rewarding LLMs to Decompose Proofs Hierarchically [29.908878832382523]
  This paper focuses on improving LLMs' ability to write proofs in formal languages that permit automated proof verification/evaluation. We work in a more natural setup where the lemmas that are directly relevant to the theorem are not given to the theorem prover at test time. We design an RL-based training algorithm that encourages the model to decompose a theorem into lemmas, prove the lemmas, and then prove the theorem by using the lemmas.
  arXiv  Detail & Related papers  (2024-11-04T05:57:40Z)
• Proof Automation with Large Language Models [6.587933406842906]
  Large Language Models (LLMs) have shown promise in automatically generating informal proofs in natural language. We propose PALM, a novel generate-then-repair approach that first prompts an LLM to generate an initial proof and then leverages targeted symbolic methods to iteratively repair low-level problems. Our results show that PALM significantly outperforms other state-of-the-art approaches, successfully proving 76.6% to 180.4% more theorems.
  arXiv  Detail & Related papers  (2024-09-22T00:19:27Z)
• DeepSeek-Prover: Advancing Theorem Proving in LLMs through Large-Scale Synthetic Data [65.5290035371111]
  We introduce an approach to generate extensive Lean 4 proof data derived from high-school and undergraduate-level mathematical competition problems. We fine-tune the DeepSeekMath 7B model on this synthetic dataset, which comprises 8 million formal statements with proofs. Our model successfully proved 5 out of 148 problems in the Lean 4 Formalized International Mathematical Olympiad (FIMO) benchmark, while GPT-4 failed to prove any.
  arXiv  Detail & Related papers  (2024-05-23T09:03:42Z)
• MUSTARD: Mastering Uniform Synthesis of Theorem and Proof Data [85.50740598523818]
  MUSTARD is a framework that masters uniform synthesis of theorem and proof data of high quality and diversity. We present a theorem-and-proof benchmark MUSTARDSAUCE with 5,866 valid data points. We perform extensive analysis and demonstrate that MUSTARD generates validated high-quality step-by-step data.
  arXiv  Detail & Related papers  (2024-02-14T05:57:58Z)
• TheoremQA: A Theorem-driven Question Answering dataset [100.39878559382694]
  GPT-4's capabilities to solve these problems are unparalleled, achieving an accuracy of 51% with Program-of-Thoughts Prompting. TheoremQA is curated by domain experts containing 800 high-quality questions covering 350 theorems.
  arXiv  Detail & Related papers  (2023-05-21T17:51:35Z)
• Generating Natural Language Proofs with Verifier-Guided Search [74.9614610172561]
  We present a novel stepwise method NLProofS (Natural Language Proof Search) NLProofS learns to generate relevant steps conditioning on the hypothesis. It achieves state-of-the-art performance on EntailmentBank and RuleTaker.
  arXiv  Detail & Related papers  (2022-05-25T02:22:30Z)
• PRover: Proof Generation for Interpretable Reasoning over Rules [81.40404921232192]
  We propose a transformer-based model that answers binary questions over rule-bases and generates the corresponding proofs. Our model learns to predict nodes and edges corresponding to proof graphs in an efficient constrained training paradigm. We conduct experiments on synthetic, hand-authored, and human-paraphrased rule-bases to show promising results for QA and proof generation.
  arXiv  Detail & Related papers  (2020-10-06T15:47:53Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.