Models · Vinci by SimpleDirect®

Open-weight AI models by Vinci.

AI models for building AI, checking security settings, and everyday tasks. Find the one that fits your work.

Explore Vinci models and research

New · Vinci MLE 1.0

Help with the work of building AI.

Vinci MLE is trained to help find out why an AI experiment went wrong, suggest a fix, or explain why nothing should change.

Evaluation coverage differs by model. Valid held-out ML-engineering and external MLE-bench evaluations were not completed for 1.0. We plan to require those evaluations before releasing MLE 1.1.

Compare Vinci MLE machine-learning engineering models

New

Open-weight release

Vinci MLE 8B 1.0

A smaller model trained to help troubleshoot AI experiments. Needs less memory than 30B.

Licence: Apache-2.0
Weight files: Full merged BF16 Safetensors · 17.58 GB

Explore Vinci MLE 8B 1.0 Model card & weights (English)

New

Open-weight release

Vinci MLE 30B 1.0

A larger model for the same work. Higher coding and tool-use test scores than 8B, but needs more memory.

Licence: Apache-2.0
Weight files: Full merged BF16 Safetensors · 57.73 GB

Explore Vinci MLE 30B 1.0 Model card & weights (English)

New

Open-weight release

Vinci MLE 123B 1.0

Our largest MLE model, built on Devstral 2, with full BF16 weights and three GGUF formats under the Mistral Modified MIT licence.

Licence: Mistral Modified MIT
Weight files: Full merged BF16 Safetensors · 250.05 GB

Explore Vinci MLE 123B 1.0 Model card & weights (English)

The download size is only the weights. Running the model needs extra memory.

Why we’re building Vinci MLE

Vinci Cyber. Defensive specialists.

Trained to review security settings and suggest fixes for reported problems. Choose a size, then check its test results and licence.

Model cards and weights on Hugging Face

Compare Vinci Cyber defensive cybersecurity models

New

Open-weight release

Vinci Cyber 8B 1.0

A smaller model trained to suggest fixes for security settings—or leave them alone when no change is needed.

Licence: Apache-2.0

Explore Vinci Cyber 8B 1.0 Model card & weights (English)

New

Open-weight release

Vinci Cyber 30B 1.0

A larger model trained for the same security work. Needs more memory than 8B.

Licence: Apache-2.0

Explore Vinci Cyber 30B 1.0 Model card & weights (English)

New

Open-weight release

Vinci Cyber 123B 1.0

Our largest Cyber model, trained to review security settings and suggest fixes.

Licence: Modified MIT (Mistral AI)

Explore Vinci Cyber 123B 1.0 Model card & weights (English)

Everyday models

For questions, writing, and conversation. Explore the Vinci family, from our first model to Piccolo and Bozza.

Historical model

flash-1-mini

Our first model, focused on legal writing and research in English and Canadian French.

Explore the story

Open-weight release

Vinci Piccolo 1.0

A small model for everyday questions, writing, and conversation.

Explore the story

Open-weight release

Vinci Bozza 1.0

A larger everyday model. See its test results and limitations before choosing it.

Explore the story

What comes next.

Our next models are in development. We’ll share more when they’re ready.

In development

Vinci Tela

The next chapter in the Vinci model family.

In development

Vinci Frontier

A new model in development for more demanding tasks.

In development

Vinci Piccolo & Bozza 1.5

Working toward better, more versatile assistants for everyday tasks.

Research beyond the mainline.

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