Pixtral-12B-2409-GGUF/README.md

---
license: apache-2.0
language:
- en
- ja
- fr
- de
- es
- it
- pt
- ru
- zh
base_model:
- mistralai/Pixtral-12B-2409
pipeline_tag: image-text-to-text
new_version: EnlistedGhost/Pixtral-12B-2409-GGUF
datasets:
- mistralai/MM-MT-Bench
tags:
- Pixtral,
- 12B,
- Vision,
- Conversational,
- Ollama,
- ggml,
- gguf,
- Image-Text-to-Text,
- Multimodal,
- MistralAI
- Llama.cpp
- Quantized
---
# Model - Pixtral-12B-2409 (GGUF + Ollama Patched) 
*[The first official MistralAI Pixtral-12B to be GGUF converted and GGUF Quantized!]*

### NEW UPDATES! (As of: November 27th, 2025) (Read below!)

-------------------------------------------------------------------------
**Description:** <br />
EnlistedGhost's GGUF/Quantized Repository of - MistralAI Pixtral-12B-2409...
This release includes GGUF+Ollama-Patched moodel files and three working projector 
(mmproj) files for the Vision Projector, offering full capabilities in Ollama or Llama.cpp.<br /><br />
**No modifications, edits, or configurations are required** to use this model with
Ollama, it works natively! **Both Vision and Text work with Ollama. (^.^)** <br />

**Personal Notes:**<br />
The quality of interaction with this model from MistralAI **has been verified
to be of a higher quality than others available.** There are other Pixtral-12B models
that are community versions, however the weights that MistralAI had are different than
from the community vertsion has and are significantly improved over community available releases.
*Thank you for taking the time to read this!<br />*

**Some explaination for this claim -** <br />
The currently available (publicly released on Huggingface) Pixtral-12B GGUF/GGUF-Quantized 
releases are from *mistral-community* and not *MistralAI* themselves.
So, I am very excited to offer the Huggingface community an Ollama and
Llama.cpp compatible version of the officially released: MistralAI/Pixtral-12B-2409
Multimodal Vision Model!<br />

**Public Notice: The statements and wording in this release DO NOT:**<br />
infer, suggest, slander or portray in a negative manner the
mistral-community and their releases! I have nothing but a very high respect
for mistral-community and their quality work. 
*(Be sure to check them out too!)* <br /> <br /> 
The wording I use is only to offer a distinction between the fact that these
GGUF/Quantized files are derrived directly from official MistralAI iterations.
*(Sorry for the long disclaimer there - I just have to be very clear that I am not
in any way saying anything negative about the already-quantized and GGUF converted
Pixtral-12B from mistral-community)*<br /><br />

Happy Inferencing! <br />
 -- Jon Z (EnlistedGhost)

---------------------------------------------

### Model Updates (As of: November 27th, 2025)
- Updated: All GGUF model file(s) with Extremely-High-Quality GGUF file(s) (You won't be disappointed!)<br />
**Final Quantized and full-BF16 modelfiles have been uploaded!!!**


## How to run this Model using Ollama
You can run this model by using the "ollama run" command.<br />
Simply copy & paste one of the commands from the list below into<br /> 
your console, terminal or power-shell window.
| Quant Type | File Size | Command |
|:-----------|:----------|:--------|
| Q2_K_S | 4.78 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q2_K_S  |
| Q2_K   | 5.12 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q2_K    |
| Q3_K_S | 5.62 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q3_K_S  |
| Q3_K_M | 6.35 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q3_K_M  |
| Q4_K_S | 7.29 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q4_K_S  |
| Q4_K_M | 7.65 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q4_K_M  |
| Q4_K_XL| 7.98 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q4_K_XL |
| Q5_K_S | 8.43 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q5_K_S  |
| Q5_K_M | 8.93 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q5_K_M  |
| Q5_K_XL| 9.14 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q5_K_XL |
| Q6_K   | 10.1 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q6_K    |
| Q6_K_M | 10.4 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q6_K_M  |
| Q6_K_XL| 11.6 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q6_K_XL |
| Q8_0   | 13.0 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:Q8_0    |
| F16    | 24.5 GB | ollama run hf.co/EnlistedGhost/Pixtral-12B-2409-GGUF:F16     |

**mmproj (Vision Projector) Files**
| Quant Type | File Size | Download-Link |
|:-----------|:----------|:--------------|
| Q8_0 | 465 MB | [[mmproj Vision Pixtral-12B rojector:Q8_0](https://huggingface.co/EnlistedGhost/Pixtral-12B-2409-GGUF/resolve/main/mmproj-Pixtral-12B-2409-Q8_0.gguf?download=true)]|
| F16 | 870 MB | [[mmproj Vision Pixtral-12B rojector:F16](https://huggingface.co/EnlistedGhost/Pixtral-12B-2409-GGUF/resolve/main/mmproj-Pixtral-12B-2409-F16.gguf?download=true)]|
| F32 | 1.74 GB | [[mmproj Vision Pixtral-12B rojector:F32](https://huggingface.co/EnlistedGhost/Pixtral-12B-2409-GGUF/resolve/main/mmproj-Pixtral-12B-2409-F32.gguf?download=true)]|

-------------------------------------------------------------------------------

### Intended Use
Same as original:
- [[MistralAI](https://huggingface.co/mistralai/Pixtral-12B-2409)]


### Out-of-Scope Use
Same as original:
- [[MistralAI](https://huggingface.co/mistralai/Pixtral-12B-2409)]


## Bias, Risks, and Limitations
Same as original:
- [[MistralAI](https://huggingface.co/mistralai/Pixtral-12B-2409)]


## Training Details
Training sets and data are from:
- [[mistralAI](https://huggingface.co/mistralai/Pixtral-12B-2409)]
<br />(This is a direct off-shoot/decendant of the above mentioned model)


## Evaluation
- This model has NOT been evaluated in any form, scope or type of method. 
- **!!! USE AT YOUR OWN RISK !!!**
- **!!! NO WARRANTY IS PROVIDED OF ANY KIND !!!**

## Citation (Original Paper)
[[MistalAI Pixtral-12B Original Paper](https://huggingface.co/papers/2410.07073)]


## Detailed Release Information
- **Originally Developed by:** [[mistralai](https://huggingface.co/mistralai/Pixtral-12B-2409)]
- **MMPROJ (Vision) Quantized by:** [[EnlistedGhost](https://huggingface.co/EnlistedGhost)]
- **Model Quantized for GGUF by:** [[EnlistedGhost](https://huggingface.co/EnlistedGhost)]
- **Modified for Ollama by:** [[EnlistedGhost](https://huggingface.co/EnlistedGhost)]
- **Model type & format:** [Quantized/GGUF]
- **License type:** [Apache-2.0]


## Attributions (Credits)
A big thank-you is extended to the below credited sources!<br />
These contributions are what made this release possible!

*Important Notice: **This is NOT a copy/paste release**, <br />
I have created unique Quantized files that were then altered further <br />
to properly work with Ollama software. <br />
This resulted in the first publicly available <br />
Pixtral-12B-2409 model that natively runs on Ollama.*

- [[MistralAI](https://huggingface.co/mistralai/Pixtral-12B-2409)]
- [[Ollama developers](https://github.com/ollama/ollama/issues/6748#issuecomment-3449009817)]

## Model Card Authors and Contact
[[EnlistedGhost](https://huggingface.co/EnlistedGhost)]