Ministral 3 14B
Language modelOpen weightsCard summarizedMinistral 3 14B is the largest model in the Ministral 3 family, offering state-of-the-art capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart.
Active (GA, v25.12)● Re-checked 12 Oct 2026 · 38 facts · 0 corrections · 1 unconfirmedSource: Ministral 3 14B Instruct 2512 model card (Hugging Face) and Introducing Mistral 3 launch post
Use it via API
Where to call it
- Mistral API (Mistral AI Studio)
ministral-14b-2512 - Self-host (weights)
mistralai/Ministral-3-14B-Instruct-2512
Good for
- Private or custom chat and assistant deployments in constrained environments
- Advanced local agentic use cases
- Fine-tuning and specialization
Overview
Ministral 3 14B
Ministral 3 14B is the largest model in the Ministral 3 family, offering state-of-the-art capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart.
- Cost
- $0.20 in
- Context
- 256K
- Input / output
- $0.20 / $0.20
- Max output
- —
- Knows up to
- —
- Released
- 2 Dec 2025
The largest Ministral 3 model, which Mistral says performs comparably to Mistral Small 3.2 24B. It pairs a 13.5B language model with a 0.4B vision encoder, has a 256K context window and is released under Apache 2.0 in Base, Instruct (FP8) and Reasoning variants. Mistral says the FP8 Instruct weights fit in 24GB of VRAM. model card intro; Key Features
ministral-14b-2512- Input / output per 1M
- $0.20 / $0.20
- Context window
- 256K
- Max output
- —
- Knows up to
- —
- Reasoning
- —
- Weights
- Open · Apache 2.0
About this card. A model card is the maker's own account of a model: what it's for, how it was trained and tested, and where it falls short. This page condenses Mistral AI's Ministral 3 14B Instruct 2512 model card (Hugging Face) and Introducing Mistral 3 launch post. Every claim carries its source page.