ministral-3-14b-instruct-2512

Mistral AI

Container

Mistral AI

ministral-3-14b-instruct-2512

Ministral 3 14B Instruct 2512 FP8 model is an efficient language model with vision capabilities, this instruct post-trained version in **FP8 precision** is fine-tuned for instruction tasks, making it ideal for chat and instruction-based use cases.

NVIDIA NIM

Layer	Label		Created
sha256:691c6e80404d56330cd1efc12b0620180d5d2a320f08d518c0165a6a44ee70ac	COPY	`./NOTICES/* /opt/nim/`	03/06/2026 7:58 AM UTC
sha256:dc80bb9208a0b8d48340bd4bc77fcd4403b366368feef91f3cd41407472becea	COPY	`./banner.txt /opt/nvidia/entrypoint.d/31-model-license.txt`	03/06/2026 7:58 AM UTC
sha256:c2572d48efc737adc1214fde49b25cfc19834c5d92b4161428205b02b90f1f2a	COPY	`./MODEL_LICENSE /opt/nim/MODEL_LICENSE`	03/06/2026 7:58 AM UTC
sha256:58939f2652b989435794278ec68d0756f3a439a29a0534a7514149b7bb76c0c4	COPY	`./VERSION /opt/nim/VERSION`	03/06/2026 7:58 AM UTC
sha256:b2ed48f2f23b464d7a04dae3cb1f8607f8459bd4985e2d8d5a239112abce9cba	COPY	`patches/yarn_scaling_rope.py /opt/nim/llm/.venv/lib/python3.12/site-packages/vllm/model_executor/layers/rotary_embedding/yarn_scaling_rope.py`	03/06/2026 7:58 AM UTC
sha256:9837e9014d95575303ab52674cc42c2df75728afc604f279d1518fb2ad5dfc2c	COPY	`patches/tool_calls.py /opt/nim/llm/.venv/lib/python3.12/site-packages/mistral_common/protocol/instruct/tool_calls.py`	03/06/2026 7:58 AM UTC
sha256:def8a9748ff26669dafd1b5d800c9862086495544d6342ac4cc86b1f1a314df8	COPY	`./model_manifest.yaml /opt/nim/etc/default/model_manifest.yaml`	03/06/2026 7:58 AM UTC
sha256:a3ed95caeb02ffe68cdd9fd84406680ae93d633cb16422d00e8a7c22955b46d4	ENV	`VLLM_LOGGING_LEVEL=ERROR`	03/06/2026 7:58 AM UTC
sha256:a3ed95caeb02ffe68cdd9fd84406680ae93d633cb16422d00e8a7c22955b46d4	ENV	`NIM_MAX_IMAGES_PER_PROMPT=10`	03/06/2026 7:58 AM UTC
sha256:a3ed95caeb02ffe68cdd9fd84406680ae93d633cb16422d00e8a7c22955b46d4	ENV	`NVIDIA_PRODUCT_NAME=NVIDIA Inference Microservice VLM NIM`	03/06/2026 7:58 AM UTC