Historical model
flash-1-mini
Where the Vinci model story began: English and Canadian French, with a focus on legal drafting and research.
Explore the storyOur story starts with flash-1-mini. Piccolo and Bozza continue it, with open weights and a visible account of what each model does—and where it falls short.
Follow the journey from a focused first model to a broader model portfolio.
Historical model
Where the Vinci model story began: English and Canadian French, with a focus on legal drafting and research.
Explore the storyOpen-weight release
A compact 4B release for everyday language work, with downloadable weights and a documented evaluation.
Explore the storyOpen-weight release
A 9B Qwen-based release that documents behavioural gains alongside instruction-following and tool-use costs.
Explore the storyTela, Cyber, and MLE are coming soon. These are development directions; model details and evaluation evidence will accompany approved releases.
Coming soon
The next chapter in the Vinci model family.
Coming soon
A specialist direction for bounded defensive cybersecurity work.
Coming soon
Machine-learning engineer: a specialist direction for ML engineering work.
Prova is our experimental Mistral-based character-training study. Its public artifacts and limitations remain part of the research record; it is not recommended for production.
Explore Prova and the research artifacts