Ministral3 / Devstral#
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.
Task |
Text Generation |
Architecture |
|
Parameters |
3B – 24B |
HF Org |
Available Models#
Ministral3:
Ministral-3-3B-Instruct-2512
Ministral-3-8B-Instruct-2512
Ministral-3-14B-Instruct-2512
Devstral:
Devstral-Small-2-24B-Instruct-2512
Architecture#
Mistral3ForConditionalGeneration
Example HF Models#
Model |
HF ID |
|---|---|
Ministral-3 3B Instruct |
|
Ministral-3 8B Instruct |
|
Ministral-3 14B Instruct |
|
Devstral Small 2 24B |
Example Recipes#
Recipe |
Description |
|---|---|
SFT — Devstral Small 2 24B on SQuAD |
|
LoRA — Devstral Small 2 24B on SQuAD |
Try with NeMo AutoModel#
1. Install (full instructions):
pip install nemo-automodel
2. Clone the repo to get the example recipes:
git clone https://github.com/NVIDIA-NeMo/Automodel.git
cd Automodel
3. Run the recipe from inside the repo:
automodel --nproc-per-node=8 examples/llm_finetune/devstral/devstral2_small_2512_squad.yaml
Run with Docker
1. Pull the container and mount a checkpoint directory:
docker run --gpus all -it --rm \
--shm-size=8g \
-v $(pwd)/checkpoints:/opt/Automodel/checkpoints \
nvcr.io/nvidia/nemo-automodel:26.02.00
2. Navigate to the AutoModel directory (where the recipes are):
cd /opt/Automodel
3. Run the recipe:
automodel --nproc-per-node=8 examples/llm_finetune/devstral/devstral2_small_2512_squad.yaml
See the Installation Guide and LLM Fine-Tuning Guide.
Fine-Tuning#
See the LLM Fine-Tuning Guide.