Stefano Zacchiroli on Nostr: last time I checked, Mistral models were merely open weight, with no training dataset ...

last time I checked, Mistral models were merely open weight, with no training dataset available nor training pipeline released as FOSS. Has that changed?