Capability Is Closing the Open-Closed Gap; Security Is Not
F5, Wednesday, August 12th, 2026
F5 Labs' August CASI evaluation adds 16 models, with Anthropic's Claude Fable topping security and capability.
F5 Labs published its August CASI evaluation covering 16 newly added models. Anthropic's claude-fable-5 debuted at the top of the board on both security and capability, while xAI's grok-4.5 debuted at the bottom.
The month's Attack Spotlight examines Morality Dilemma, a jailbreak that hands a model a corrupted ethical rule and asks it to reason from it.
The accompanying news analysis looks at an open-closed capability gap, arguing that while capability is closing the gap between open and closed models, security is not. The article is by Lee Ennis with contributions from Malcolm Heath.