PromptArmor Study Says Open-Weight Models Are 88% More Susceptible to Indirect Prompt-Injection Attacks
Open-weight models 88% more susceptible to cyber attacks
AI BriefingIn an evaluation of 9,800 trials, PromptArmor found that open-weight models (DeepSeek, Kimi, GLM, and Qwen) had a 26.7% susceptibility rate to real-world indirect prompt-injection attack chains, compared with 14.2% for closed-source models (GPT, Claude, and Gemini); open-weight models were 88% more susceptible.
Source: PromptArmor: Threat Intelligence · promptarmor.com
Event follow-up1
View all event coverage- PromptArmor Study Says Open-Weight Models Are 88% More Susceptible to Indirect Prompt-Injection AttacksPromptArmor: Threat Intelligence · 2 days ago