announcement aidev SIG 4/5

Mistral releases Shieldstral, a 3B open-weights multimodal safety classifier

Mistral released Shieldstral, a 3-billion-parameter open-weights multimodal safety classifier that it claims beats models up to seven times its size. Relevant to anyone running content moderation or guardrails locally, though the source gives no benchmark names or license terms.

PUBLISHED2026-08-04
OBSERVED2026-08-11
AGE21d
SOURCES1
  • New model: Shieldstral, safety classifier
  • 3B parameters, open weights
  • Multimodal input
  • Claimed to outperform models up to 7x its size (no benchmarks named in source)
new-modelsecuritycapability

COMMUNITY

No curated reactions recorded for this event. Facts and takes are kept in separate layers — community context is added by hand, never blended into the record above.