Get the app
← back to catalog · registered 2026-09-02 09:03

EZCon/Qwen2.5-VL-3B-Instruct-abliterated-mlx

EZCon 3.8B multimodal second-order
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/EZCon%2FQwen2.5-VL-3B-Instruct-abliterated-mlx"
Response includes
  • classification m1
  • files 16
  • benchmarks 11 entries
  • hub_downloads_all_time 129
  • author_summary 31 models
  • readme_text full
Try next
GET/models?author=EZCon
All 31 models by EZCon
 GET/methods/m1
What "m1" means · taxonomy
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
How we label → About M1 →
Refusal direction extracted via
Extraction technique

Difference-of-means

FromArditi et al. (2024) ↗
Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
What is a refusal direction? → Read the paper ↗
Downloads · lifetime
129
109 last 30d - active
Likes
0
Model age
6w ago
created 2026-09-02
Downloads over time
Now190→from0↑0%
0701392090 on Sep 2190 on Oct 17190 on Oct 15SepOct
Sep 2 → Oct 17 · 46 snapshots · spans 45 days

Benchmarks

Benchmark Score Source
Entertainment 0.8 UGI
Hazardous 1.8 UGI
Natural Intelligence 9.42 UGI
Political lean -41.7% UGI
Sensitive-Info 9.4 UGI
SocPol 0.5 UGI
UGI 27.1 UGI
Willingness (10) 6.2 UGI
W10-Adherence 4.5 UGI
W10-Direct 8 UGI
Writing 17.47 UGI

Genealogy 0 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.

huihui-ai/Qwen2.5-VL-3B-Instruct-abliterated↓ 19,988 this lineage
EZCon/Qwen2.5-VL-3B-Instruct-abliterated-mlx↓ 109

Metadata

Languages
en
Tags
mlx safetensors qwen2_5_vl multimodal abliterated uncensored image-text-to-text conversational en base_model:huihui-ai/Qwen2.5-VL-3B-Instruct-abliterated base_model:finetune:huihui-ai/Qwen2.5-VL-3B-Instruct-abliterated region:us

Related

Other by EZCon
Huihui-Qwen3-VL-4B-Instruct-abliterated-mlx 449 Huihui-gemma-4-E4B-it-abliterated-4bit-mlx 419 Qwen2-VL-2B-Instruct-abliterated-mlx 387 Huihui-Qwen3-VL-4B-Instruct-abliterated-8bit-mlx 326 Huihui-Qwen3-VL-2B-Instruct-abliterated-mlx 307
Same base model
mradermacher/Qwen2.5-VL-3B-Instruct-abliterated-GGUF 3K mradermacher/Qwen2.5-VL-3B-Instruct-abliterated-i1-GGUF 2K
Total size
6.99 GB
Files
16
Quantizations
1
Registered
2026-09-02 09:03
Last updated on HF
2026-09-02 08:16
Source huggingface.co/EZCon/Qwen2.5-VL-3B-Instruct-abliterated-mlx ↗ Base model huihui-ai/Qwen2.5-VL-3B-Instruct-abliterated →

Files by quantization

Auxiliary files 16 files 7.01 GB
model-00001-of-00002.safetensors 4.99 GB aafb4e60 download
model-00002-of-00002.safetensors 2.00 GB a4806c01 download
tokenizer.json 10.9 MB 3fd16973 download
vocab.json 2.65 MB 4783fe10 download
model.safetensors.index.json 75.8 KB b61aaba0 download
config.json 1.87 KB 4af1fee8 download
.gitattributes 1.53 KB 52373fe2 download
processor_config.json 1.05 KB 2243a037 download
chat_template.json 1.03 KB 732bd68b download
chat_template.jinja 1017 B 6c226632 download
README.md 911 B efada6e4 download
tokenizer_config.json 737 B dded43f4 download
special_tokens_map.json 644 B 3a784031 download
added_tokens.json 629 B 06135f3c download
preprocessor_config.json 351 B 8273ab1b download
generation_config.json 229 B 9b492590 download

README current version from Hugging Face

license_name: qwen-research
license_link: https://huggingface.co/Qwen/Qwen2.5-VL-3B-Instruct/blob/main/LICENSE
language:

  • en
    pipeline_tag: image-text-to-text
    tags:
  • multimodal
  • abliterated
  • uncensored
  • mlx
    library_name: mlx
    base_model: huihui-ai/Qwen2.5-VL-3B-Instruct-abliterated

EZCon/Qwen2.5-VL-3B-Instruct-abliterated-mlx

This model was converted to MLX format from huihui-ai/Qwen2.5-VL-3B-Instruct-abliterated
using mlx-vlm version 0.6.17.
Refer to the original model card for more details on the model.

Use with mlx

pip install -U mlx-vlm

python -m mlx_vlm.generate --model EZCon/Qwen2.5-VL-3B-Instruct-abliterated-mlx --max-tokens 100 --temperature 0.0 --prompt "Describe this image." --image <path_to_image>

README history 1 version

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

  1. 2026-09-02Upload folder using huggingface_hub0fd2a54911 B
    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