OpenRT: An Open-Source Red Teaming Framework for Multimodal LLMs

TL;DR

OpenRT is a comprehensive, modular framework for systematically evaluating the safety vulnerabilities of multimodal large language models through diverse attack strategies and extensive empirical testing.

Contribution

We introduce OpenRT, a scalable, modular red-teaming framework that standardizes attack interfaces and integrates 37 attack methods for thorough MLLM safety evaluation.

Findings

Frontier models show safety gaps with up to 49.14% attack success rate.

Reasoning models do not inherently have better robustness against multi-turn jailbreaks.

Models fail to generalize safety defenses across different attack paradigms.

Abstract

The rapid integration of Multimodal Large Language Models (MLLMs) into critical applications is increasingly hindered by persistent safety vulnerabilities. However, existing red-teaming benchmarks are often fragmented, limited to single-turn text interactions, and lack the scalability required for systematic evaluation. To address this, we introduce OpenRT, a unified, modular, and high-throughput red-teaming framework designed for comprehensive MLLM safety evaluation. At its core, OpenRT architects a paradigm shift in automated red-teaming by introducing an adversarial kernel that enables modular separation across five critical dimensions: model integration, dataset management, attack strategies, judging methods, and evaluation metrics. By standardizing attack interfaces, it decouples adversarial logic from a high-throughput asynchronous runtime, enabling systematic scaling across diverse models. Our framework integrates 37 diverse attack methodologies, spanning white-box gradients, multi-modal perturbations, and sophisticated multi-agent evolutionary strategies. Through an extensive empirical study on 20 advanced models (including GPT-5.2, Claude 4.5, and Gemini 3 Pro), we expose critical safety gaps: even frontier models fail to generalize across attack paradigms, with leading models exhibiting average Attack Success Rates as high as 49.14%. Notably, our findings reveal that reasoning models do not inherently possess superior robustness against complex, multi-turn jailbreaks. By open-sourcing OpenRT, we provide a sustainable, extensible, and continuously maintained infrastructure that accelerates the development and standardization of AI safety.