Claimant scorecard · AERS v2.1 · Calibrating
Mistral
3 claims tracked in the Responsibility Ledger. 3 pending grades.
AERS
—
Insufficient closed grades
Pending
3
Open horizons
Closed
0
Graded outcomes
First tracked
Sep 9, 2026
Open horizons
Mistral: Open weights for Large 4 ship by October 31, after API-only preview for cybersecurity partners and state authorities
Mistral: Launched Large 4 October 6 via API at 1.05 trillion parameters, promised open-weights release by end of October 2026
Mistral: Raised €3 billion at €21 billion valuation, with CEO pledging to double owned compute in five years
Invalidator — If Mistral's owned capacity fails to grow significantly, or if the company shifts back to renting most compute, or if ARR growth stalls below trajectory needed to justify €21B valuation.
About this scorecard
The AI Execution Risk Score (AERS) is a 0-100 metric quantifying the gap between Mistral’s public AI claims and demonstrated delivery. Higher AERS = stronger track record. Each claim above is drawn from a primary source linked in the original Ledger entry; the horizon date is when the claim becomes graded under the published methodology. Materiality is the editor’s assessment of the claim’s formality from 1 (PR statement) to 5 (earnings call or SEC filing).
AERS v2.1 · Methodology in active calibration · Not investment advice.