Mistral AI introduces Shieldstral open safety classifier
Mistral AI has released Shieldstral, a 3B open-weights multimodal safety classifier available under the Apache 2.0 license. The model accepts plain-language policies at inference time to evaluate text and images, returning a calibrated safety score from a single forward pass without requiring model retraining.
1 min readai safetycontent moderationmistral ai