Mistral's Shieldstral: 3B open-weights model for multimodal moderation

Mistral released Shieldstral, a 3-billion-parameter open-weights multimodal safety classifier that adapts to different content policies at inference without retraining. It outperforms larger models on text safety and sets new standards for multimodal moderation, running efficiently on a single 16GB NVIDIA GPU.