We present StepFun-Prover Preview, a large language model designed for formal theorem proving through tool-integrated reasoning. Using a reinforcement learning pipeline that incorporates tool-based interactions, StepFun-Prover can achieve strong performance in generating Lean 4 proofs with minimal sampling. Our approach enables the model to emulate human-like problem-solving strategies by iteratively refining proofs based on real-time environment feedback. On the miniF2F-test benchmark, StepFun-Prover achieves a pass@1 success rate of 70.0%. Beyond advancing benchmark performance, we introduce an end-to-end training framework for developing tool-integrated reasoning models, offering a promising direction for automated theorem proving and Math AI assistant.
@article{arxiv.2507.20199,
title = {StepFun-Prover Preview: Let's Think and Verify Step by Step},
author = {Shijie Shang and Ruosi Wan and Yue Peng and Yutong Wu and Xiong-hui Chen and Jie Yan and Xiangyu Zhang},
journal= {arXiv preprint arXiv:2507.20199},
year = {2025}
}