{"acceptedBy":{"actor":"ai","name":null,"on":"2026-10-22"},"aliases":[],"change":{"last4":0,"percent":null,"prior4":0},"description":{"editedBy":null,"state":"none"},"href":"/wiki/model-shieldstral/","itemCount":2,"key":"model--shieldstral","kind":"model","name":"Shieldstral","neighbours":[{"key":"licence--apache-2-0","kind":"licence","name":"Apache-2.0","relation":"licensed_under","weight":2},{"key":"licence--apache-2-0","kind":"licence","name":"Apache-2.0","relation":"mentioned_with","weight":2},{"key":"task--content-moderation","kind":"task","name":"content moderation","relation":"mentioned_with","weight":2},{"key":"maker--mistral-ai","kind":"maker","name":"Mistral AI","relation":"made_by","weight":1},{"key":"concept--vram","kind":"concept","name":"VRAM","relation":"mentioned_with","weight":1},{"key":"maker--mistral-ai","kind":"maker","name":"Mistral AI","relation":"mentioned_with","weight":1},{"key":"maker--openai","kind":"maker","name":"OpenAI","relation":"mentioned_with","weight":1},{"key":"model--gpt-oss-safeguard","kind":"model","name":"gpt-oss-safeguard","relation":"mentioned_with","weight":1},{"key":"model--gpt-oss-safeguard-20b","kind":"model","name":"GPT-OSS-Safeguard-20B","relation":"mentioned_with","weight":1},{"key":"model--llavaguard","kind":"model","name":"LlavaGuard","relation":"mentioned_with","weight":1},{"key":"model--llavaguard-7b","kind":"model","name":"LlavaGuard-7B","relation":"mentioned_with","weight":1},{"key":"task--multimodal-safety-classification","kind":"task","name":"multimodal safety classification","relation":"mentioned_with","weight":1},{"key":"task--policy-based-classification","kind":"task","name":"policy-based classification","relation":"mentioned_with","weight":1}],"relatedItems":[{"publishedAt":"2026-07-16T00:00:00Z","reason":"model--shieldstral","title":"Mistral AI releases Shieldstral 1.0, a 3B multimodal safety classifier under Apache 2.0","url":"/decisions/mistral-shieldstral-1-0-3b/"},{"publishedAt":"2025-10-29T00:00:00Z","reason":"model--shieldstral","title":"OpenAI releases gpt-oss-safeguard, open-weight safety reasoning models that read your policy","url":"/decisions/openai-gpt-oss-safeguard/"}],"schemaVersion":1,"series26":[0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0],"series8":[0,0,0,0,0,0,0,0],"weeksOfHistory":52}