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

Shiftedx/qwopus3.6-35b-a3b-coder-abliterated-mxfp4-vision-mlx

Shiftedx 35B MoE 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/Shiftedx%2Fqwopus3.6-35b-a3b-coder-abliterated-mxfp4-vision-mlx"
Response includes
  • classification m1
  • files 15
  • hub_downloads_all_time 601
  • author_summary 24 models
  • readme_text full
Try next
GET/models?author=Shiftedx
All 24 models by Shiftedx
 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
601
127 last 30d - stable
Likes
1
Model age
3mo ago
created 2026-07-01
Downloads over time
Now641→from45↑1,324%
1524447270145 on Jul 1641 on Oct 11641 on Oct 10JulAugSepOct
Jul 1 → Oct 11 · 54 snapshots · spans 102 days

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.

Jackrong/Qwopus3.6-35B-A3B-Coder this lineage
Shiftedx/qwopus3.6-35b-a3b-coder-abliterated-mxfp4-vision-mlx↓ 127

Metadata

License
apache-2.0
Tags
mlx safetensors qwen3_5_moe mlx-vlm mxfp4 lm-studio apple-silicon qwen3_6 coder agent tool-use function-calling

Related

Other by Shiftedx
qwen3.8-27b-aeon-ultimate-uncensored-attention8-bf16recurrence-vision-mtplx 1K qwen3.8-27b-aeon-ultimate-uncensored-mxfp4-vision-mtplx 1K Qwen3.8-27B-AEON-ULTIMATE-UNCENSORED-MLX-MIXED-VISION-MTPLX 543 qwopus3.8-27b-flash-abliterated-mxfp4-vision-mtplx 541 ornith-1.5-35b-a3b-abliterated-attention8-bf16recurrence-vision-mtplx 474
Same base model
Shiftedx/qwopus3.6-35b-a3b-coder-abliterated-mxfp8-vision-mlx 58
Total size
18.0 GB
Files
15
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-07-05 23:07
Source huggingface.co/Shiftedx/qwopus3.6-35b-a3b-coder-abliterated-mxfp4-vision-mlx ↗ Base model Jackrong/Qwopus3.6-35B-A3B-Coder ↗

Files by quantization

Auxiliary files 15 files 18.0 GB
model-00003-of-00004.safetensors 5.00 GB 520a8e3f download
model-00002-of-00004.safetensors 5.00 GB ec3dc4b9 download
model-00001-of-00004.safetensors 4.98 GB 8e1450f4 download
model-00004-of-00004.safetensors 2.18 GB 7f3f3434 download
model-vision-00001-of-00001.safetensors 852 MB d4579339 download
tokenizer.json 19.1 MB 87a7830d download
model.safetensors.index.json 161 KB 97fa6085 download
config.json 48.5 KB 896c3900 download
tokenizer_config.json 15.0 KB dc7d3fb6 download
chat_template.jinja 7.70 KB 68a73673 download
README.md 3.09 KB 84b6e8d8 download
.gitattributes 1.53 KB 52373fe2 download
processor_config.json 991 B 8f29fe38 download
preprocessor_config.json 390 B 2ea84a43 download
video_preprocessor_config.json 385 B 3ba673a5 download

README current version from Hugging Face

license: apache-2.0
base_model: Jackrong/Qwopus3.6-35B-A3B-Coder
base_model_relation: quantized
library_name: mlx
pipeline_tag: image-text-to-text
tags:

  • mlx
  • mlx-vlm
  • mxfp4
  • safetensors
  • lm-studio
  • apple-silicon
  • qwen3_5_moe
  • qwen3_6
  • coder
  • agent
  • tool-use
  • function-calling
  • vision
  • thinking-off
  • abliterated

qwopus3.6-35b-a3b-coder-abliterated-mxfp4-vision-mlx

Vision-enabled MLX MXFP4 conversion of Jackrong/Qwopus3.6-35B-A3B-Coder, prepared by Shiftedx for Apple Silicon, MLX, and LM Studio.

Build Notes

  • Quantized primary language weights with mxfp4 at group size 32.
  • Kept MoE router and gate modules in affine 8-bit group size 64 for compatibility.
  • Added Qwen3.5/Qwen3.6-MoE-compatible vision components and validated image grounding locally.
  • Removed source MTP tensors and set MTP/next-token prediction layer counts to 0 for LM Studio compatibility.
  • Set tool_parser_type to qwen3_coder.
  • Patched the chat template so enable_thinking defaults to false when the runtime honors that template variable.
  • Applied a research refusal-direction weight edit using residual-direction orthogonalization.

Local Validation

Hardware reference: Apple M4 Max Apple Silicon with 64 GB unified memory.

Validated locally on June 30, 2026 and July 1, 2026 with LM Studio and direct MLX/VLM loading.

Check Result
LM Studio load Passed at 32k context, parallel 1, GPU max.
Basic text completion Passed; answered 2+2 with 4 and stopped.
Code completion Passed; produced a simple valid add(a, b) function.
Direct MLX/VLM image color smoke Passed; answered Red.
Direct MLX/VLM OCR smoke Passed; answered FABLE 42.
LM Studio OpenAI-compatible image smoke Passed; answered Red and FABLE 42.
LM Studio native image smoke Passed; answered Red and FABLE 42.
Thinking-off behavior Smoke checks returned 0 reasoning tokens.
LM Studio logs No warnings, errors, tracebacks, KV-cache issues, or tokenizer-regex warnings in the validation window.

This is a smoke-validated release, not a full benchmark suite. Broader downstream evaluation is still recommended for production use.

Recommended LM Studio Defaults

After downloading in LM Studio, load the model by repo name:

lms load shiftedx/qwopus3.6-35b-a3b-coder-abliterated-mxfp4-vision-mlx --context-length 32768 --parallel 1 --gpu max

Recommended profile defaults:

  • Preset/template: Qwen3 thinking-compatible Jinja template with <|im_end|> stop.
  • Thinking: off by default through the included chat template.
  • Context length: 200000 when memory allows; 32768 was used for local validation.
  • Sampling: temperature 0.6, top-k 20, top-p 0.95, min-p enabled at 0.
  • Repeat penalty: off by default.
  • Load: parallel 1, GPU max.

Provenance

README history 2 versions

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

  1. 2026-07-05Add validation hardware reference6b3ebc43.1 KB
    Loading...
  2. 2026-07-01Add files using upload-large-folder toole560b5f3 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