Score: 2

DafnyPro: LLM-Assisted Automated Verification for Dafny Programs

Published: January 8, 2026 | arXiv ID: 2601.05385v1

By: Debangshu Banerjee, Olivier Bouissou, Stefan Zetzsche

BigTech Affiliations: Amazon

Potential Business Impact:

Helps computers prove code is correct.

Business Areas:
Developer Tools Software

We present DafnyPro, an inference-time framework that enhances LLMs for generating verification annotations in Dafny. DafnyPro comprises three key components: a diff-checker that prevents modifications to base program logic, a pruner that removes unnecessary invariants, and a hint-augmentation system that retrieves and applies predefined, problem-independent proof strategies. We evaluate DafnyPro using Claude Sonnet 3.5 and 3.7 on four benchmarks: Clover, MBPP-Dafny, HumanEval-Dafny, and DafnyBench, achieving consistent performance gains in all cases. Notably, on DafnyBench, the most challenging benchmark, Claude Sonnet 3.5 enhanced with DafnyPro achieves 86% correct proofs, a 16 pp improvement over the base model. We also fine-tune two Qwen models on training data derived from verification attempts by larger models enhanced with DafnyPro. Our 7B and 14B models achieve 68% and 70% correct proofs on DafnyBench, respectively, demonstrating that smaller models can maintain high verification accuracy.

Country of Origin
πŸ‡ΊπŸ‡Έ United States

Page Count
14 pages

Category
Computer Science:
Software Engineering