mistral
12B parameters
Commercial OK
Multimodal

Pixtral 12B

Mistral's multimodal entry. 12B parameters, vision + text, Apache 2.0. Good document and chart understanding.

License: Apache 2.0·Released Sep 17, 2024·Context: 131,072 tokens

Overview

Mistral's multimodal entry. 12B parameters, vision + text, Apache 2.0. Good document and chart understanding.

Strengths

  • Apache 2.0
  • Multimodal
  • 12B fits 16GB

Weaknesses

  • Vision quality below Qwen-VL family

Quantization variants

Each quantization trades model quality for file size and VRAM. Q4_K_M is the most popular starting point.

QuantizationFile sizeVRAM required
Q4_K_M7.0 GB10 GB
Q8_013.0 GB16 GB

Get the model

Ollama

One-line install

ollama run pixtral:12bRead our Ollama review →

HuggingFace

Original weights

huggingface.co/mistralai/Pixtral-12B-2409

Source repository — direct quantization required.

Hardware that runs this

Cards with enough VRAM for at least one quantization of Pixtral 12B.

Compare alternatives

Models worth comparing

Same parameter band, plus what's one tier above and below — so you can decide what actually fits your hardware.

Frequently asked

What's the minimum VRAM to run Pixtral 12B?

10GB of VRAM is enough to run Pixtral 12B at the Q4_K_M quantization (file size 7.0 GB). Higher-quality quantizations need more.

Can I use Pixtral 12B commercially?

Yes — Pixtral 12B ships under the Apache 2.0, which permits commercial use. Always read the license text before deployment.

What's the context length of Pixtral 12B?

Pixtral 12B supports a context window of 131,072 tokens (about 131K).

How do I install Pixtral 12B with Ollama?

Run `ollama pull pixtral:12b` to download, then `ollama run pixtral:12b` to start a chat session. The default quantization is Q4_K_M.

Does Pixtral 12B support images?

Yes — Pixtral 12B is multimodal and accepts text + vision inputs. Vision support requires a runner that handles its image-conditioning architecture.

Source: huggingface.co/mistralai/Pixtral-12B-2409

Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify model claims.