SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2603.24465 · 2026

MechMath: Sorrifier-Driven Formal Decomposition Workflow for Automated Theorem Proving

Xiao-Shan Gao, Junqi Liu, Lihong Zhi, Ruichen Qiu, Yichuan Cao, Dakai Guo, Ruyong Feng, Sklms

arXiv · PDF

Code that ran

15 of 15 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 3 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 11 failures of our instrument; 0 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2603.24465.json · badge: /api/ran/2603.24465.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
MechMath/MechMath-v1 canonical (classifier label)
classifier basis: census
15 ran · 0 no recorded run
FunctionStatusWhere it lives
extract_content_by_tag Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/utils.py
pointer only (licence: NONE) · get_code("87b518b0a65eb03c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

extract_error Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/lean/feedback.py
pointer only (licence: NONE) · get_code("a110ead50bbb0182")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

extract_error_dict Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/utils.py
pointer only (licence: NONE) · get_code("c7f66af130c3a829")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

extract_from_box Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/utils.py
pointer only (licence: NONE) · get_code("a50b9e4b0f25862b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

extract_logic_error Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/lean/feedback.py
pointer only (licence: NONE) · get_code("bf3107ef589c2680")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

filter_error_messages Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/lean/sorrifier.py
pointer only (licence: NONE) · get_code("701260b6fa0fa8e1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

format_feedback_messages Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/lean/feedback.py
pointer only (licence: NONE) · get_code("1747a79cf49add0e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

gen_one_sorry_of_block Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/lean/sorrifier.py
pointer only (licence: NONE) · get_code("b3370e77a671c397")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_indent Ran
ran, honoured the contract we drafted
MechMath/MechMath-v1/src/mechmath/lean/parser.py
pointer only (licence: NONE) · get_code("17b8e8719543609f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_search_result Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/lean/search.py
pointer only (licence: NONE) · get_code("00df48c980211da4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

is_error Ran
ran, violated the contract we drafted
MechMath/MechMath-v1/src/mechmath/lean/client.py
pointer only (licence: NONE) · get_code("af0fa648e07de9b6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

leandex Ran
ran, honoured the contract we drafted
MechMath/MechMath-v1/src/mechmath/lean/search.py
pointer only (licence: NONE) · get_code("29c7bc71b7ad2700")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

loogle Ran
ran, honoured the contract we drafted
MechMath/MechMath-v1/src/mechmath/lean/search.py
pointer only (licence: NONE) · get_code("abf345a2fa0aa735")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

remove_comments Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/lean/code_blocks.py
pointer only (licence: NONE) · get_code("5569d923e1ca9d92")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

remove_sorry_statements Ran
ran; our contract draft was wrong, not the code
MechMath/MechMath-v1/src/mechmath/lean/code_blocks.py
pointer only (licence: NONE) · get_code("57422d1a4865d73c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MechMath/MechMath-v1 when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 14 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
remove_sorry_statements 1 other paper
1 dated earlier
2505.03171 (2025-05)
CombiBench: Benchmarking LLM Capability for Combinatorial Mathematics

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Recent advances in large language models (LLMs) and LLM-based agents have substantially improved the capabilities of automated theorem proving. However, for problems that require complex mathematical reasoning, current systems seldom succeed in their initial attempt, necessitating iterative adjustments to their proof strategies. Existing approaches for handling failed attempts typically either iteratively fix errors within the proof or discard the entire proof and regenerate it from scratch. The former leads to progressively longer contexts, which degrade the model's ability to attend to the remaining unresolved subproblems, while the latter is inefficient, as it may abandon mostly correct reasoning due to localized errors. To address this dilemma, we present MechMath, an agent system centered on a Sorrifier-driven formal decomposition paradigm. By leveraging the sorry placeholder in Lean to precisely isolate unresolved subgoals while preserving the surrounding verified proof structure, MechMath extracts each failed subproblem into a clean, self-contained context and resolves it independently. This avoids both the waste of full regeneration and the excessive context length induced by repeated repairs. Experimental results on challenging mathematical competition benchmarks, including IMO 2025, Putnam 2025, miniF2F, and a subset of ProverBench, demonstrate that our agent achieves significant advantages in proving efficiency. Mathematical reasoning has long been one of the core research objectives in artificial intelligence. In recent years, with the rapid advancement of large language models (LLMs), their ability to handle complex mathematical reasoning tasks at the natural language level has improved significantly. State-of-the-art systems have achieved gold-medal performance in high school mathematics competitions (e.g., IMO 2025), while systems such as Deep-Mind's Aletheia (Feng et al., 2026) are progressively approaching the frontier of tackling research-level problems. Despite these advances, theorem proving in natural languages is still plagued by verification difficulties, leading to potential logical hallucinations and erroneous conclusions, and it lacks scalable automated validation (Chen et al., 2025b;a). In contrast, formal languages such as Lean, Rocq, and Isabelle offer superior logical rigor and verifiability. In formal math, previous neural theorem proving research primarily focused on training a single model * Equal contribution. † Corresponding author.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2603.24465")
get_code_for_paper("2603.24465")
have("2603.24465")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2603.24465.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2603.24465.svg)](https://syntology.ai/paper/2603.24465)

Connect an agent — have() is free.