Can I run Ministral 3 14B locally?

Yes — the weights are open under Apache 2.0 and available on Hugging Face. The FP8 instruct model fits in 24 GB of VRAM, and quantised versions will run on less.

What does it cost via API, and how long is the context?

Mistral prices it at $0.20 per million tokens for both input and output, with a 262,144-token context window.

How does it compare to Mistral Small 3.2 24B?

Mistral says performance is comparable despite the smaller parameter count, making the 14B the better choice when you need to stay within tighter memory or cost constraints.

Ministral 3 14B 2512

5506°

#llm #mistralai #vision #agents #multilingual #long context

Run this model

Offers / Ads

OpenRouter Terminal / API

Pay per token

Chat with this model in the browser or call it from your code with one API key.

Mistral | 262,144 token context | $0.20/M input, $0.20/M output. The largest model in the Ministral 3 family, Ministral 3 14B offers frontier capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart. A powerful and efficient language model with vision capabilities.

Guides for this model

Ollama vs LM Studio (2026): which local AI app to use, and which top models each one runs

How much VRAM do I actually need for local AI?

What is a GGUF file? Quantized models explained (Q4, Q8, fp8, NF4)

All guides

Captain's notes

Written by Captain, the house editor

Notes from 2026-10-13

Ministral 3 14B 2512 is Mistral's largest edge-optimised model, pairing a 13.5B language model with a 0.4B vision encoder to deliver performance Mistral claims is comparable to its Mistral Small 3.2 24B. It ships under Apache 2.0, fits in 24 GB of VRAM at FP8, and brings native function calling, a 256 k-token context and multilingual support to hardware that would choke on bigger models.

Known for

Fits in 24 GB VRAM at FP8; quantise further for tighter hardware
Native function calling and JSON output for agentic pipelines
Vision encoder handles image analysis alongside text
256 k-token context window for long documents
Apache 2.0 licence: commercial use, fine-tuning, modification all permitted

Watch out for

FP8 weights only in the instruct version; base models in the family use BF16
Vision quality degrades with very wide or narrow images — crop to near 1:1 before sending
Mistral recommends temperature below 0.1 for production; higher values need careful testing

Best for

Prompts to try

Agentic tool-use plan

You have two tools: search_web(query: str) and send_email(to: str, subject: str, body: str). A user asks: 'Find the current CEO of Mistral AI and send me a summary email at user@example.com.' Write the exact sequence of tool calls you would make, with arguments, and explain your reasoning between each step.

Describe an attached image

Attach a photo of any product, document or scene. Describe what you see in detail: identify the main subject, read any visible text, note colours and layout, and suggest two ways this image could be used in a marketing context.

Multilingual contract clause

Translate the following clause into French, Spanish, German and Japanese, keeping legal register intact. Then flag any term that may not have a direct legal equivalent in one of those languages and suggest a local alternative: 'The licensor grants a non-exclusive, royalty-free, worldwide licence to use, reproduce and distribute the software in source or binary form.'

Compare with

Mistral Large 4
API-only model
1,048,576 token context · $0.68/M input, $2.09/M output
Mistral Medium 3.5
API-only model
262,144 token context · $1.50/M input, $7.50/M output
Mistral Large 3 2512
API-only model
262,144 token context · $0.50/M input, $1.50/M output
Mistral Large 2407
API-only model
131,072 token context · $2.00/M input, $6.00/M output

More text models from mistralai

Questions people ask

Can I run Ministral 3 14B locally?: Yes — the weights are open under Apache 2.0 and available on Hugging Face. The FP8 instruct model fits in 24 GB of VRAM, and quantised versions will run on less.
What does it cost via API, and how long is the context?: Mistral prices it at $0.20 per million tokens for both input and output, with a 262,144-token context window.
How does it compare to Mistral Small 3.2 24B?: Mistral says performance is comparable despite the smaller parameter count, making the 14B the better choice when you need to stay within tighter memory or cost constraints.

Captain's verdict
A rare edge model that takes vision, long context and agents seriously — worth running locally if you have a 24 GB card.

Model card on Hugging Face

Captain writes these notes from the model's own description and model card. Prices and context sizes are the provider's at the time of writing and can change.

Ranking badge

Updated Oct 22, 2026

Made this model? Show this badge on your page.

Paste it into your Hugging Face card or Civitai description.

Change colors / personalize / get code

Not this creator?

Get your own badge

How this score adds up

This model has no popularity data from any source yet, so it carries a stand-in number below 7,000° that keeps the lists complete. It never ranks above a model with data. The breakdown appears here as soon as a source reports it.

Glossary: the terms used here

Heat score: One number for how much a model is used and wanted right now, from 0° to 9,999°. Every list on this site that says hottest is sorted by it and nothing else.
Floor: The 7,000° every model with real data starts from. The scale runs from 0° to 9,999°; the part below the floor is reserved for stand-ins (models without data), so that a model with data always ranks above one without. The floor is the same for everyone and decides nothing; the degrees above it are the measurement.
Download-equivalents: Everything the score treats as a download, added up: downloads themselves, each torrent seeder as 1,000, each AI Horde image and each run on a generation site as one.
Log scale: Counting by tens, not by ones: going from 1,000 to 10,000 downloads adds the same degrees as going from 10,000 to 100,000. It keeps giants from crowding out everything else.
Likes: Thumbs up and likes from every source, worth half a download each on the log scale.
Adoption rate: How many downloads a day a model gets right now, from the difference between two readings of a source. A quarter of the weight of the download term.
Measured and estimated: Measured means two readings of a source exist and the difference is real. Estimated means only today's total is known and the past is spread over the model's life; the site says which one it is showing.
Seeder: Someone who keeps a copy of the weights online for others to download over a torrent, or a hosted mirror doing the same. Each one counts as 1,000 downloads.
AI Horde: A volunteer network that generates images for free on donated computers. The number of images it has made with a model is public, and each counts as one download.
Generation sites: Sites that host their own copies of a model and let people generate with it, such as PixAI, SeaArt and Yodayo. Their downloads count like Civitai's and each generation run as one download.
OpenRouter tokens: A text model served through an API, such as Claude or GPT, has no weights to download, so its usage is measured in tokens: the pieces of text it read and wrote through OpenRouter, a gateway many apps use, over the last 30 days. Every 10,000,000 tokens count as one download. Source: OpenRouter (openrouter.ai/rankings), CC BY 4.0.
Full model (checkpoint): The weights themselves, the file that generates on its own. It gets a bonus because add-ons depend on it.
LoRA and embedding: Small add-on files that steer someone else's weights toward a style, a character or a concept. They cannot run alone.
Family and weights: The base a model is built on, such as SD 1.5, SDXL, Flux or Wan. Every model inherits the release date of its family's weights.
Age: Degrees taken off for how long ago the family's weights came out: 150° a year, up to 750°. It keeps old downloads that had years to pile up from freezing the top of the list.
Family boost: A small, bounded nudge for LoRAs of a family this site is pushing forward right now, always listed in the breakdown.
Stand-in score: A stable number below the floor given to a model no source reports on yet, so lists stay complete. It never ranks above a model with data.
New! (grace period): A model less than 7 days old shows New! instead of a score and holds no place in the rankings. Its score is computed from day one, but a model one day old grows faster than anything, so without the wait the week's #1 would always be whatever came out yesterday rather than what held up for a week.
Trend: How a model's growth over a period compares with the typical model's growth over the same period. Faster than typical reads as a rise, slower as a fall.
Typical model: The middle of the catalog: half the models with a figure grew faster over the period, half slower.

Similar Models

AI Models for Image, Video & Text Generation

Ministral 3 14B 2512 is a text model for the Text / LLM family listed on ModelTrends.ai, a read-only catalog of open source AI image, video and text models. Compare it with other Text / LLM models and check its heat score to see how it is trending.

More Text / LLM models llm mistralai vision agents multilingual long context local

Browse by Model Family