GGUF quantizations of
coder3101/Ministral-3-14B-Reasoning-2512-heretic for use with llama.cpp and compatible tools.
This is a fine-tuned version of Mistral's Ministral-3-14B-Reasoning-2512 vision-language model. It supports:
The model includes a custom chat template with reasoning support. The format uses:
1llama-cli -m Ministral-3-14B-Reasoning-2512-heretic-Q4_K_M.gguf \
2 -p "[INST]What is the capital of France?[/INST]" \
3 -n 256
1llama-mtmd-cli \
2 -m Ministral-3-14B-Reasoning-2512-heretic-Q4_K_M.gguf \
3 --mmproj Ministral-3-14B-Reasoning-2512-heretic-mmproj-bf16.gguf \
4 -p "Describe this image in detail." \
5 --image /path/to/image.jpg
1llama-server \
2 -m Ministral-3-14B-Reasoning-2512-heretic-Q4_K_M.gguf \
3 --mmproj Ministral-3-14B-Reasoning-2512-heretic-mmproj-bf16.gguf \
4 --port 8080
1curl http://localhost:8080/v1/chat/completions \
2 -H "Content-Type: application/json" \
3 -d '{"model": "ministral", "messages": [{"role": "user", "content": "What is 2+2?"}]}'
