Multi-turn attack success rates run 2x to 10x higher than single-turn baselines across eight open-weight LLMs in an earlier evaluation.
| Publisher | Cisco |
| Report | Proprietary Problems: How Frontier Closed Models Collapse Under Iterative Pressure |
| Published | 27 May 2026 |
| Topics | Model Robustness, Adversarial Attacks, Open Models |
Published by Cisco in Proprietary Problems: How Frontier Closed Models Collapse Under Iterative Pressure, 27 May 2026. The figure is taken from the report as published; the full methodology is in the source.