#Mistral models’

Enkrypt AI’s report reveals serious safety flaws in Mistral’s vision-language models that enable generation of harmful content, urging continuous security improvements in multimodal AI systems.