# Oedipus: LLM-enchanced Reasoning CAPTCHA Solver

arXiv: https://arxiv.org/abs/2405.07496 · HTML: https://mlsanity.com/paper/2405.07496/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 13 May 2024.

## Built on by (most cited first, 22 of 22)

- [Frontier AI's Impact on the Cybersecurity Landscape](https://mlsanity.com/paper/2504.05408/) (2025)
- [MCA-Bench: A Multimodal Benchmark for Evaluating CAPTCHA Robustness Against VLM-based Attacks](https://mlsanity.com/paper/2506.05982/) (2025)
- [AgentBay: A Hybrid Interaction Sandbox for Seamless Human-AI Intervention in Agentic Systems](https://mlsanity.com/paper/2512.04367/) (2025)
- [Systematic Categorization, Construction and Evaluation of New Attacks against Multi-modal Mobile GUI Agents](https://mlsanity.com/paper/2407.09295/) (2024)
- [Reasoning under Vision: Understanding Visual-Spatial Cognition in Vision-Language Models for CAPTCHA](https://mlsanity.com/paper/2510.06067/) (2025)
- [CAPTURE: A Benchmark and Evaluation for LVLMs in CAPTCHA Resolving](https://mlsanity.com/paper/2512.11323/) (2025)
- [Spatial CAPTCHA: Generatively Benchmarking Spatial Reasoning for Human-Machine Differentiation](https://mlsanity.com/paper/2510.03863/) (2025)
- [Next-Gen CAPTCHAs: Leveraging the Cognitive Gap for Scalable and Diverse GUI-Agent Defense](https://mlsanity.com/paper/2602.09012/) (2026)
- [Broken Gates: Re-evaluating Web Bot Defenses in the Age of LLM Agents](https://mlsanity.com/paper/2607.18659/) (2026)
- [Throttling Web Agents Using Reasoning Gates](https://mlsanity.com/paper/2509.01619/) (2025)
- [Beyond BeautifulSoup: Benchmarking LLM-Powered Web Scraping for Everyday Users](https://mlsanity.com/paper/2601.06301/) (2026)
- [VIPER Strike: Defeating Visual Reasoning CAPTCHAs via Structured Vision-Language Inference](https://mlsanity.com/paper/2601.06461/) (2026)
- [Hallucination-Resistant Security Planning with a Large Language Model](https://mlsanity.com/paper/2602.05279/) (2026)
- [CaptchaMind: Training CAPTCHA Solvers via Reinforcement Learning with Explicit Reasoning Supervision](https://mlsanity.com/paper/2605.19538/) (2026)
- [HLL: Can Agents Cross Humanity's Last Line of Verification?](https://mlsanity.com/paper/2606.02449/) (2026)
- [Rethinking Side-Channel Analysis: Automated Discovery and Analysis of Side-Channel Leakage with LLM-Assisted Agents](https://mlsanity.com/paper/2605.17406/) (2026)
- [On the Internet, Nobody Knows You're an LLM Bot: Unmasking Web Agents with Multi-Layer Fingerprinting](https://mlsanity.com/paper/2606.30119/) (2026)
- [Breaking Darknet CAPTCHAs with general purpose LLM](https://mlsanity.com/paper/2608.28794/) (2026)
- [CAPTCHAs in the Agentic Era: Solvers That Learn from Every Encounter](https://mlsanity.com/paper/2609.02393/) (2026)
- [CaptchaArena: A Large-Scale, Fine-Grained Dataset for Training Computer-Use Agents on Interactive CAPTCHAs](https://mlsanity.com/paper/2609.31957/) (2026)
- [CaptchaAgent: Verifiable Visual Reasoning for CAPTCHA Solving](https://mlsanity.com/paper/2610.12866/) (2026)
- [DarKnight: Retrieval-Guided Agentic Framework for Breaking Dark Web CAPTCHA Gates](https://mlsanity.com/paper/2610.19830/) (2026)

## Abstract (arXiv)

CAPTCHAs have become a ubiquitous tool in safeguarding applications from automated bots. Over time, the arms race between CAPTCHA development and evasion techniques has led to increasingly sophisticated and diverse designs. The latest iteration, reasoning CAPTCHAs, exploits tasks that are intuitively simple for humans but challenging for conventional AI technologies, thereby enhancing security measures. Driven by the evolving AI capabilities, particularly the advancements in Large Language Models (LLMs), we investigate the potential of multimodal LLMs to solve modern reasoning CAPTCHAs. Our empirical analysis reveals that, despite their advanced reasoning capabilities, LLMs struggle to solve these CAPTCHAs effectively. In response, we introduce Oedipus, an innovative end-to-end framework for automated reasoning CAPTCHA solving. Central to this framework is a novel strategy that dissects the complex and human-easy-AI-hard tasks into a sequence of simpler and AI-easy steps. This is achieved through the development of a Domain Specific Language (DSL) for CAPTCHAs that guides LLMs in generating actionable sub-steps for each CAPTCHA challenge. The DSL is customized to ensure that each unit operation is a highly solvable subtask revealed in our previous empirical study. These sub-steps are then tackled sequentially using the Chain-of-Thought (CoT) methodology. Our evaluation shows that Oedipus effectively resolves the studied CAPTCHAs, achieving an average success rate of 63.5%. Remarkably, it also shows adaptability to the most recent CAPTCHA designs introduced in late 2023, which are not included in our initial study. This prompts a discussion on future strategies for designing reasoning CAPTCHAs that can effectively counter advanced AI solutions.