Introducing Shieldstral. | Mistral
https://mistral.ai/news/shieldstral/
Shieldstral — Mistral safety publication; watched for whether further documentation is added alongside it
2
Lines added
3
Lines removed
Was read
2026-10-06
2a2689a2d8cb3c47…
Changed by
2026-11-04
478e421e29a8b5a3…
2 lines added, 3 lines removed
Left column is the version read 2026-10-06, right is 2026-11-04. Unchanged stretches are elided. Both bodies are kept in full and are addressed by the hashes above.
⋯ 65 unchanged lines
6666 Shieldstral is a step toward moderation that adapts to context instead of forcing every product through one frozen taxonomy. We're continuing to push on multilingual coverage, longer-document robustness, and broader multimodal safety — and we'd love to see what the community builds on top of it.
6767 BTW, we're hiring! If you want to help make AI better, see our careers page .
6868
6969 Shieldstral
70−
71−Apache 2.0
70+New
7271
73−Text and image safety classification against a supplied policy.
72+Compact multimodal moderation model for text and image safety classification.
7473 Text-to-text
7574 Image-to-Text
7675
7776 Read documentation