Mistral AI Releases Shieldstral 1.0 3B: An Open-Weights Policy-Adaptive Multimodal Safety Classifier Matching Models 7× Its Size

Mistral AI launched Shieldstral 1.0 3B, an open-weight multimodal safety classifier that adapts to different content policies via plain-language queries, matching the performance of much larger models. It runs efficiently on a single GPU with 16GB VRAM, supports commercial use under Apache 2.0, and suits diverse industries needing flexible, local content moderation.