Not published
Only provider-documented reasoning controls are shown; missing controls remain unpublished rather than inferred.
MODEL REFERENCE / MISTRAL AI
Shieldstral 1.0
verified model reference.
Compact multimodal moderation model for prompt/response safety classification, refusal detection and policy filtering.
32,768context tokens
Not publishedmax output
2026-10-07last verified
PRIMARY-SOURCE VERIFIED
Shieldstral 1.0 at a glance.
Mistral AI documents a 32,768-token context window. Input modalities: text + image. Output: classification. Pricing status: No direct Standard price published.
MODEL IDshieldstral-1.0
CONTEXT WINDOW32,768tokens
MAX OUTPUTNot published
KNOWLEDGE CUTOFFNot published
STANDARD RATENot publishedper 1 million tokens
OUTPUT RATE—per 1 million tokens
CAPABILITY CONTRACT
Published model behavior.
text + image
Input modalities come from the cited official model documentation.
classification
Output modality and output-limit claims remain separate so an undocumented token cap is not guessed.
PRICING STATUS
No direct Standard price published.
No direct Standard price published.
SXF preserves the provider-published native billing basis. Only calculator-eligible generative token models enter token-cost arithmetic; specialist units and input-only embeddings remain visible without forced conversion.
CAVEATS
What the source record does not assume.
01
Public Preview model released August 4, 2026 under Apache 2.0.
02
No directly comparable hosted Standard inference rate is asserted.
EXPLORE
Keep the model in context.
OFFICIAL SOURCES
Verification links.
WHAT CHANGED
Shieldstral 1.0: what changed.
No post-baseline factual changes recorded for Shieldstral 1.0 since 2026-10-07.
VERIFIED BASELINE2026-10-07Current facts remain aligned with the SXF ledger.
VERIFIED CHANGE HISTORY
Shieldstral 1.0 verified history.
Shieldstral 1.0
Baseline verified: 32,768 context · Not published max output
Official evidence ↗Append-only SXF ledger · 1 verified event · chain head 8e73a9cb15ea…