Can LLMs Deobfuscate Binary Code? A Systematic Analysis of Large Language Models into Pseudocode Deobfuscation
2026-04-09 • Software Engineering
Software Engineering
AI summaryⓘ
The authors studied how well large language models (LLMs) can understand tricky, hidden code in computer programs. They created a test called BinDeObfBench to check how good these models are at dealing with different hiding techniques used at various stages of program creation. Their results showed that being good at reasoning and having specific training for the task helped more than just using bigger models. They also found that models trained this way could handle tough hiding tricks and work across different computer systems. This work provides a tool and insights for improving future methods in uncovering hidden code.
binary deobfuscationlarge language modelsreverse engineeringobfuscationfine-tuninginstruction set architecturepre-compilationcompile-time optimizationpost-compilationin-context learning
Authors
Li Hu, Xiuwei Shang, Jieke Shi, Shaoyin Cheng, Junqi Zhang, Gangyang Li, Zhou Yang, Weiming Zhang, David Lo
Abstract
Deobfuscating binary code remains a fundamental challenge in reverse engineering, as obfuscation is widely used to hinder analysis and conceal program logic. Although large language models (LLMs) have shown promise in recovering semantics from obfuscated binaries, a systematic evaluation of their effectiveness is still lacking. In this work, we present BinDeObfBench, the first comprehensive benchmark for assessing LLM-based binary deobfuscation across diverse transformations spanning pre-compilation, compile-time, and post-compilation stages. Our evaluation shows that deobfuscation performance depends more on reasoning capability and domain expertise than on model scale, and that task-specific supervised fine-tuning consistently outperforms broad domain pre-training. Reasoning models can maintain robustness under severe obfuscation, generalize across different instruction set architectures (ISAs) and optimization levels. In-context learning benefits standard models but yields limited gains for reasoning models. Overall, our study highlights the importance of task-specific fine-tuning and reasoning-driven strategies, and positions BinDeObfBench as a basis for future work in binary deobfuscation.