Get the app
← back to catalog · registered 2026-08-22 13:56

ccharnkij/Ministral-3-14B-Reasoning-2512-Uncensored

ccharnkij Mistral 14B multimodal
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model. 

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/ccharnkij%2FMinistral-3-14B-Reasoning-2512-Uncensored"
Response includes
  • classification m-uncensored
  • files 15
  • benchmarks 11 entries
  • hub_downloads_all_time 139
  • author_summary 20 models
  • readme_text full
Try next
GET/models?author=ccharnkij
All 20 models by ccharnkij
 GET/models?family=mistral&method=m-uncensored
Sibling mistral models · same method
 GET/methods/m-uncensored
What "m-uncensored" means · taxonomy
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M-U
Primary method

Uncensored (method unknown)

No other method signals detected in this model.
Confidence
LOW
Why this label 3 signals
Weak or ambiguous signals. Best guess based on catalog patterns; treat as tentative and check the evidence below.
  • 'uncensored' in name/tags but no 'abliterated' marker
  • method not identifiable from author declaration alone
  • may be DPO fine-tune, prompt engineering, or unknown technique
How we label → About M-U →
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · lifetime
139
47 last 30d - stable
Likes
1
Descendants
2
in 2 direct forks
Model age
2mo ago
created 2026-07-16
Downloads over time
Now161→from24↑571%
177012217524 on Jul 15161 on Oct 11JulAugSepOct
Jul 15 → Oct 11 · 53 snapshots · spans 88 days

Benchmarks

Portrait before abliteration
Benchmarks of the base model as it stood before the refusal-removal operation. Compare with the numbers above to see what the operation cost.
Benchmark Score Source
Entertainment 1.2 UGI
Hazardous 1.8 UGI
Natural Intelligence 16.35 UGI
Political lean -16.7% UGI
Sensitive-Info 11.99 UGI
SocPol 0.8 UGI
UGI 32.16 UGI
Willingness (10) 7.2 UGI
W10-Adherence 8.5 UGI
W10-Direct 6 UGI
Writing 25.63 UGI

Genealogy 2 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

mistralai/Ministral-3-14B-Reasoning-2512 this lineage
ccharnkij/Ministral-3-14B-Reasoning-2512-Uncensored↓ 47 2 forks
mradermacher/Ministral-3-14B-Reasoning-2512-Uncensored-GGUF↓ 1,573
mradermacher/Ministral-3-14B-Reasoning-2512-Uncensored-i1-GGUF↓ 1,251

Variants by this author 2 formats · 2K downloads combined

The same weights this author released in different packaging. Pick the format that matches your runtime.

Metadata

License
apache-2.0
Languages
en
Tags
safetensors mistral3 mistral ministral uncensored text-generation image-text-to-text multimodal reasoning lora conversational en

Related

Other by ccharnkij
Qwen3.5-9B-Uncensored-GGUF 3K Ministral-3-14B-Reasoning-2512-Uncensored-GGUF 2K gpt-oss-120b-Uncensored-GGUF 636 Gemma-4-E4B-it-Uncensored-GGUF 555 gemma-4-31B-it-Uncensored-GGUF 546
Same base model
ccharnkij/Ministral-3-14B-Reasoning-2512-Uncensored-GGUF 2K
Total size
26.0 GB
Files
15
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-07-16 04:07
Source huggingface.co/ccharnkij/Ministral-3-14B-Reasoning-2512-Uncensored ↗ Base model mistralai/Ministral-3-14B-Reasoning-2512 ↗

Files by quantization

Auxiliary files 15 files 26.0 GB
model-00001-of-00006.safetensors 4.59 GB e22f3f51 download
model-00003-of-00006.safetensors 4.53 GB 8579aedc download
model-00004-of-00006.safetensors 4.53 GB d0f091fd download
model-00005-of-00006.safetensors 4.53 GB c96b3742 download
model-00002-of-00006.safetensors 4.53 GB bfaeab3b download
model-00006-of-00006.safetensors 3.26 GB d905986b download
tokenizer.json 16.3 MB 57757562 download
tokenizer_config.json 200 KB e1f27c1d download
model.safetensors.index.json 56.6 KB c2462f3d download
chat_template.jinja 6.06 KB beb4d612 download
README.md 2.51 KB 20369f0b download
config.json 2.12 KB 4df52d22 download
.gitattributes 1.53 KB 52373fe2 download
processor_config.json 701 B 4c628661 download
generation_config.json 136 B b2315368 download

README current version from Hugging Face

license: apache-2.0
base_model: mistralai/Ministral-3-14B-Reasoning-2512
tags:

  • mistral
  • ministral
  • uncensored
  • text-generation
  • image-text-to-text
  • multimodal
  • reasoning
  • lora
    language:
  • en
    pipeline_tag: image-text-to-text

Ministral-3-14B-Reasoning-2512-Uncensored

A fine-tuned version of Ministral-3-14B-Reasoning-2512 with uncensored responses. This model removes the default safety filters and restrictions, allowing for more open-ended and unrestricted conversations including adult content.

Model Details

  • Base Model: mistralai/Ministral-3-14B-Reasoning-2512
  • Fine-tuning Method: LoRA (Low-Rank Adaptation)
  • Model Format: Safetensors
  • Modalities: Text + Images
  • License: Apache 2.0

What's Different

  • Removed refusal behaviors and safety filters
  • Capable of generating adult content
  • Responds to prompts the base model would typically decline
  • Retains the full reasoning and multimodal capabilities of Ministral-3-14B

Reasoning Mode

This model is a reasoning model and thinks before answering. It uses Mistral's custom chat template for proper formatting. Make sure to use the included chat_template.jinja when running inference for best results.
Non-thinking also works as well, if non-thinking system prompt is provided.

GGUF Version

A quantized GGUF version of this model is also available for use with llama.cpp, Ollama, and text-generation-webui:
👉 Ministral-3-14B-Reasoning-2512-Uncensored-GGUF

⚠️ Known behavior: refusal can happen if it's already part of the context

If a safe thinking pattern appears during normal conversation (e.g., a stray "As a safe AI assistant..." style response), this could poison the context resulting in refusal in the later turns. If you get an
unexpected refusal:

  • Start a fresh conversation/context rather than continuing in the same thread

💬 Feedback

If you use this model, I'd love to hear from you! Feel free to:

  • Leave a comment on this model page
  • Open a discussion for questions or suggestions
  • Report any issues or unexpected behaviors

Your feedback helps improve future versions.

⚠️ Disclaimer

This model is intended for research and educational purposes only.

License

This model is licensed under the Apache 2.0 License.

README history 1 version

The author's README evolved over time. Click a version to see its content at that point.

  1. 2026-07-16Upload folder using huggingface_hubed816772.4 KB
    Loading...
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Get the app Browse the catalog
Open in Abliteration