Model CoverageLarge Language ModelsMistral AIMinistral-3-3B-Instruct-2512

Ministral-3-3B-Instruct-2512

View as Markdown

Ministral is Mistral AI’s efficient small model series optimized for on-device and edge use cases. Devstral is a code-focused model built on the same architecture, designed for software engineering agents.

Both use the Mistral3ForConditionalGeneration architecture.

Set up NeMo AutoModel with the latest container or follow the installation instructions.

Fine-Tune Ministral-3-3B-Instruct-2512

From the repository root, run:

uv run automodel examples/llm_finetune/mistral/ministral3_3b_squad.yaml \
  --nproc-per-node 8

Model Reference

Model Architecture

PropertyValue
TaskText Generation
ArchitectureMistral3ForConditionalGeneration
Parameters3B
Hugging Face Organizationmistralai

Available Models

ModelHF ID
Ministral-3 3B Instructmistralai/Ministral-3-3B-Instruct-2512