ministral-3:14b
Upstream reference: Ministral 3 14B Instruct 2512
Publisher: Mistral AI
Mistral documents this instruction-tuned Ministral 3 checkpoint with a language component and separate vision encoder.
Publisher-described context, separate from Sigma5C observations and recorded catalog statistics.
Published upstream specifications
- Language component
- 13.5 billion parameters Source
- Vision encoder
- 0.4 billion parameters Source
- Published precision
- FP8 Source
- Published context
- 256k tokens, publisher notation Source
- Upstream license label
- Apache-2.0 Source
Publisher-described uses
- Text interaction, visual understanding and tool-oriented workflows described by Mistral. Source
Reference limitations
A recorded tag and an upstream card association do not verify the deployed checkpoint, quantization, configured context or current hosted availability.
Publisher sources
- Ministral 3 14B Instruct 2512 official documentation — Mistral AI. Checked .
Links
Recorded references do not establish downloadable weights, licensing or current availability.
13.9B
parameters
Q4_K_M
quantization
mistral3
architecture
262144
context
apache-2.0
license
2025-12-02
released
Where ministral-3:14b comes from
Company / lab
Mistral AI
See every model and family from this maker.
Model family
ministral-3
Other versions and sizes in this line.
completiontoolsvision