MerLean

AI-assisted workflows for Lean 4 autoformalization, theorem proving, and machine-checked verification.

About MerLean

MerLean is a Lean-based verifier and prover for mathematics and code. As a verifier, it turns mathematical solutions, programs, and AI-generated code into machine-checkable certificates that help certify correctness and catch mistakes; as a prover, MerLean-Prover searches for formal proofs for hard and open problems, and supports vibe coding without hallucination by grounding generated code in formal checks.

21/179 Top 5/6
Lean Eval Benchmark
10/23 Top 1
FormalQualBench
12/12, 789 mins Top 1
Putnam 2025 Slice

Fundraising, Advisors & Cofounders

We are speaking with pre-seed investors who are interested in capital-efficient assurance infrastructure for AI-generated critical code, formal verification, and machine-checkable scientific and software work.

Pre-seed investors

We are looking for partners who understand deep technical products, early research-to-product transitions, and high-trust markets where correctness evidence matters.

Advisors and cofounder-level collaborators

MerLean already has a working proof stack. We are selectively adding advisors and cofounder-level technical collaborators who can make the company stronger in formal verification and LLM post-training for formal reasoning.

MerLean-Prover Results

MerLean-Prover uses three agent types around a recursive proof-plan loop: planning, focused artifact checking, and Lean compile / repair. The proof assistant remains the judge; each closed proof passes the Lean kernel and axiom audit described in the paper.

Lean Eval top submitters
Rank Submitter Solved
1Seed Prover (ByteDance)62
2Aleph Prover(logicalintelligence.com)56
3Aristotle (Harmonic)52
4Stealth Model42
5Antigravity (Multi-Model Ensemble: Gemini 3.1 Pro, Gemini 3 Flash, Claude 4.6 Sonnet/Opus)21
6MerLean-Prover21
7Claude Opus 4.7 (1M context)17
8GPT-5.516
9EVO (deepthought.com.au)3
10GPT-5.5 Codex2

Source: Lean Eval public leaderboard, fetched June 9, 2026. Ranked by main benchmark problems solved; internal test problems do not count toward the score.

On the public Lean Eval Benchmark, MerLean-Prover is currently in the top #5/#6 tier among 25 models: 21 solved problems, tied by solved count with the #5 entry, from a small academic team competing on the same table as large company-backed Lean agents.

FormalQualBench closure rate
System Solved / 23
MerLean-Prover (ours) 10
OpenGauss8
Aristotle*6
Claude Code (Skills)5
Codex5
opencode (Opus)5
Claude Code4
Claude Code (MCP)3
Codex (Skills + MCP)3

† Nine solves close within the 4-hour budget; one extended run closes in 4h40m. *Aristotle results are reported but unvalidated in the benchmark source.

Putnam 2025 wall-clock time (minutes)
Prob. Arist. Seed 1.5 Axiom NLA Ours
A130601109727
A260301803031
A3301201654444
A418024010716938
A5----5182040235
A6602402598925
B115054027055161
B2253606514255
B34030433016
B4--12011230853
B54202402548843
B618018049479761
Solved10/1211/1212/1212/1212/12
Total----25773889789

Comparison columns are reproduced from Numina-Lean-Agent Table 3. Bold cells in the paper mark fastest solves; here they are highlighted.

Cite

@article{ren2026merlean, title = {MerLean: An Agentic Framework for Autoformalization in Quantum Computation}, author = {Ren, Yuanjie and Li, Jinzheng and Qi, Yidi}, journal = {arXiv preprint arXiv:2602.16554}, year = {2026}, eprint = {2602.16554}, archivePrefix = {arXiv}, primaryClass = {cs.LO}, doi = {10.48550/arXiv.2602.16554} } @misc{li2026merleanproverrecursiveloopingharness, title = {MerLean-Prover: A Recursive Looping Harness for End-to-End Lean 4 Theorem Proving}, author = {Jinzheng Li and Zeru Zhu and Yuanjie Ren}, year = {2026}, eprint = {2605.26959}, archivePrefix = {arXiv}, primaryClass = {cs.LO}, url = {https://arxiv.org/abs/2605.26959} }

Publications

Research papers from the MerLean project.

arXiv 2026
MerLean-Prover: A Recursive Looping Harness for End-to-End Lean 4 Theorem Proving
Jinzheng Li*, Zeru Zhu*, Yuanjie Ren*
arXiv preprint arXiv:2605.26959 · cs.LO

* Equal contribution.

arXiv 2026
MerLean: An Agentic Framework for Autoformalization in Quantum Computation
Yuanjie Ren*, Jinzheng Li*, Yidi Qi*
arXiv preprint arXiv:2602.16554 · cs.LO

Videos

Demos and presentations showcasing MerLean in action.

MerLean Product Demo (1080p)

A three-minute walkthrough of the MerLean workflow, from research artifact ingestion to machine-checked Lean output and interactive exploration. Open the video directly.

Blog

Updates on autoformalization, Lean 4, and quantum computation.

Examples

Browse the formalization output produced by MerLean.

Fault-Tolerant Quantum Computation

Formalization of Low-overhead fault-tolerant quantum computation by gauging logical operators. Covers stabilizer formalism, Pauli algebra, transversal gates, gauging graphs, and fault-tolerant measurement — 1,450 Lean declarations from 48 statements.

Balanced Product Quantum Codes

Formalization of Balanced Product Quantum Codes. Covers chain complexes, cohomology, CSS & LDPC codes, expander graphs, Cayley graphs, and the LPS construction.