Instructions to use osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF",
	filename="mmproj-Qwopus3.6-27B-v2-abliterated-F16.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": [
				{
					"type": "text",
					"text": "Describe this image in one sentence."
				},
				{
					"type": "image_url",
					"image_url": {
						"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
					}
				}
			]
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16
# Run inference directly in the terminal:
llama-cli -hf osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16
# Run inference directly in the terminal:
llama-cli -hf osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16
# Run inference directly in the terminal:
./llama-cli -hf osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16

Use Docker

docker model run hf.co/osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16

LM Studio
Jan

vLLM

How to use osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16

Ollama
How to use osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF with Ollama:
```
ollama run hf.co/osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16
```

Unsloth Studio

How to use osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF to start chatting

How to use osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF with Docker Model Runner:
```
docker model run hf.co/osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16
```

Lemonade

How to use osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF:F16

Run and chat with the model

lemonade run user.osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF-F16

List all available models

lemonade list

osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF

✅ MULTIMODAL. Bundled mmproj.gguf (~928 MB, F16) preserves the full Qwen3.6-VL vision tower. Use it with llama-server --mmproj or llama-mtmd-cli for text + image inference.

⚠️ Custom fork required. Native TQ3_1S inference needs the turbo-tan/llama.cpp-tq3 fork — stock llama.cpp will fail to load. Build it with cmake -B build -DGGML_METAL=ON -DCMAKE_BUILD_TYPE=Release && cmake --build build -j.

ℹ️ TQ3_1S CLI exposure. The runtime supports TQ3_1S (LLAMA_FTYPE_MOSTLY_TQ3_1S = 43), but the official fork's tools/quantize/quantize.cpp did not list it in the QUANT_OPTIONS table. To produce this file we added one row: { "TQ3_1S", LLAMA_FTYPE_MOSTLY_TQ3_1S, " 4.00 bpw TurboQuant two-scale" }. The quantizer kernels (quantize_tq3_1s) and runtime ops are unchanged.

TQ3_1S (TurboQuant two-scale, 4.00 BPW Walsh-Hadamard) of a ZeroFuse-abliterated Qwopus 3.6 27B v2 (the Jackrong Claude-Opus reasoning distill of Qwen 3.6 27B). Refusals reduced from 91/100 → 4/100 with KL drift of just 0.0176. By the osmAPI research team and TERV.Pro student research team.

⚡ TL;DR

Property	Value
Disk size	~14 GB (13 GB LM + 928 MB mmproj)
BPW	4.00 (TQ3_1S, effective ~3.5 bpw via Walsh-Hadamard transform)
Scheme	TurboQuant TQ3_1S — Walsh-Hadamard-transform weight format with two half-block scales per 32-weight block (lower-overhead variant of TQ3_4S, slightly higher perplexity). By turbo-tan.
Refusal rate (ZeroFuse, n=100)	4/100 (vs vanilla Qwopus 91/100)
KL divergence vs vanilla (at BF16)	0.0176
Vision	✅ via paired `mmproj.gguf`
Recommended RAM/VRAM	16 GB+ Apple Silicon / 12 GB GPU (lower with `-ctk q4_0 -ctv tq3_0`)
Runtime	REQUIRES turbo-tan/llama.cpp-tq3 fork. Stock llama.cpp will NOT load this file. The official `llama-quantize` in that fork did not expose `TQ3_1S` in its CLI table; we patched it in (one-line addition).
Released by	osmAPI · TERV.Pro

🎚️ All osmQwopus variants

The full osmQwopus family from osmAPI — same ZeroFuse-abliterated weights (refusal 4/100, KL 0.0176), different quant schemes for different runtimes.

Quant	Format	BPW	Disk	Vision	Runtime	Link
8-bit	MLX	8.50	~27 GB	✅ native	mlx-vlm	`…-8-bit-mlx`
6-bit	MLX	6.66	~21 GB	✅ native	mlx-vlm	`…-6-bit-mlx`
OptiQ 3.7bpw	MLX	~3.7	~14 GB	✅ ViT spliced	mlx-vlm	`…-OptiQ-3.7bpw-mlx`
Q8_0	GGUF	8.50	~28 GB	✅ via mmproj	llama.cpp	`…-8-bit-GGUF`
Q6_K	GGUF	~6.56	~22 GB	✅ via mmproj	llama.cpp	`…-6-bit-GGUF`
Q4_K_M	GGUF	~4.92	~16 GB	✅ via mmproj	llama.cpp	`…-Q4_K_M-GGUF`
TQ3_4S	GGUF	4.00 (~3.5 eff)	~14 GB	✅ via mmproj	llama.cpp-tq3	`…-TQ3_4s-GGUF`
TQ3_1S (this repo)	GGUF	4.00 (~3.5 eff)	~14 GB	✅ via mmproj	llama.cpp-tq3	— (you are here)

👉 All variants share the same abliterated base weights — pick by your runtime (Apple Silicon → MLX; CUDA/CPU/cross-platform → GGUF) and your RAM budget.

🧬 Lineage

Qwen/Qwen3.6-27B                              (Qwen Team — base multimodal pretrain)
        │
        ▼
Jackrong/Qwopus3.6-27B-v2                     (Jackrong — Claude-Opus reasoning distill)
        │
        ▼
ZeroFuse abliteration (TPE-50)          (osmAPI · TERV.Pro)
   ├── 25 random startup trials
   ├── 2 community priors (coder3101, wangzhang)
   └── 23 TPE smart-sampling trials → best at trial 45
        │
        ▼
HF safetensors → F16 GGUF via llama.cpp-tq3   (osmAPI · TERV.Pro)
        │
        ▼
this repo — osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1S GGUF + paired mmproj.gguf

Direct upstream links:

🏛️ Foundation: Qwen/Qwen3.6-27B
🎓 Claude-Opus distill: Jackrong/Qwopus3.6-27B-v2
🔓 Abliteration tool: ZeroFuse by osmAPI
🧮 Quantization tool: turbo-tan/llama.cpp-tq3 (a fork of ggml-org/llama.cpp)

📊 Abliteration Results

Stage	Refusals (n=100) ↓	KL divergence ↓
Vanilla Jackrong/Qwopus3.6-27B-v2	91 / 100	— (reference)
Community prior: coder3101 (T27)	4 / 100	0.0359
Community prior: wangzhang (T28)	30 / 100	0.0259
TPE best (T45) — shipped here	4 / 100	0.0176
TPE second-best (T37)	5 / 100	0.0210

→ 96% reduction in refusals with capability preserved (KL ≈ 0.018, well below the 0.3 healing threshold). No SFT / LoRA healing was required.

🧪 Method (TPE-50 with community priors → llama.cpp GGUF)

Step 1. Abliteration (ZeroFuse TPE-50, BF16 source)

25 random startup trials + 2 community priors enqueued (coder3101 dir=37.97, wangzhang dir=34.66) + 23 TPE smart-sampling trials.
Best Pareto trial: T45 (direction_index=41.42) — 4/100 refusals at KL=0.0176.
Auto-saved via ZeroFuse's LoRA-adapter merge path with vision tower fully intact.

Total ZeroFuse wall-clock: ~13 h on M4 Max 128 GB.

Step 2. HF safetensors → F16 GGUF

python convert_hf_to_gguf.py \
    /path/to/Qwopus3.6-27B-v2-abliterated \
    --outfile Qwopus3.6-27B-v2-abliterated-F16.gguf \
    --outtype f16

The turbo-tan fork's converter registers Qwen3_5ForConditionalGeneration natively and emits proper SSM tensors (ssm_a, ssm_conv1d, ssm_alpha, ssm_beta, ssm_out) alongside the gated-attention layers.

Step 3. Vision tower → mmproj.gguf

python convert_hf_to_gguf.py \
    /path/to/Qwopus3.6-27B-v2-abliterated \
    --outfile mmproj-Qwopus3.6-27B-v2-abliterated-F16.gguf \
    --outtype f16 \
    --mmproj

This emits a separate 928 MB GGUF containing the 27-block Qwen3-VL ViT (334 vision tensors at F16/F32) plus the multimodal projector.

Step 4. Quantization

# After patching tools/quantize/quantize.cpp to add TQ3_1S to QUANT_OPTIONS:
./build/bin/llama-quantize --pure \
  Qwopus3.6-27B-v2-abliterated-F16.gguf \
  osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1S.gguf \
  TQ3_1S 16

📦 Use it

llama-server (OpenAI-compatible HTTP, multimodal)

./build/bin/llama-server \
  -m osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1S.gguf \
  --mmproj mmproj-Qwopus3.6-27B-v2-abliterated-F16.gguf \
  --host 127.0.0.1 --port 8080 \
  -ngl 99 -c 8192 -np 1 \
  -ctk q4_0 -ctv tq3_0 -fa on \
  --jinja --no-cache-prompt --cache-ram 0

Then point any OpenAI-compatible client at http://127.0.0.1:8080/v1.

llama-mtmd-cli (one-shot multimodal generation)

./build/bin/llama-mtmd-cli \
  -m osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1S.gguf \
  --mmproj mmproj-Qwopus3.6-27B-v2-abliterated-F16.gguf \
  --image photo.jpg \
  -p "Describe this image briefly."

llama-cli (text-only)

./build/bin/llama-cli \
  -m osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1S.gguf \
  -ngl 99 \
  -c 8192 \
  --jinja \
  -p "Explain the difference between SSM and softmax attention in three sentences."

Ollama / LM Studio / Jan

Drop the two GGUF files into the runtime's models directory; standard ⓘ multimodal flow.

🧪 Quantization details

Source weights: BF16 abliterated checkpoint (12 shards, ~50 GB) — ZeroFuse T45 merged into Jackrong/Qwopus3.6-27B-v2.
Intermediate: F16 GGUF (53.8 GB, 851 tensors) produced by convert_hf_to_gguf.py from turbo-tan/llama.cpp-tq3.
Final quantization: see Step 4 above.
Vision projector: F16, 928 MB, shipped as mmproj-Qwopus3.6-27B-v2-abliterated-F16.gguf in this repo. Mandatory for image input; standard llama.cpp --mmproj flag.

Architecture notes

Qwen 3.6 27B uses a hybrid attention stack — 3 GatedDeltaNet (linear attention / SSM) layers followed by 1 full-softmax-attention layer, repeated 16× for 64 total layers; hidden 5120, vocab 248320, context 262144. The hybrid arch is supported in the turbo-tan/llama.cpp-tq3 fork (the upstream Qwen3_5ForConditionalGeneration registration). The SSM kernels run via llama.cpp's ssm_* tensor types.

⚠️ Behavior caveats

Uncensored. Refusal directions were surgically removed; this model will answer prompts the parent would refuse. Use responsibly and within applicable law. The release is provided for safety research, red-teaming, and creative/educational use cases.
Multimodal preserved. Pair the LM GGUF with mmproj.gguf (in this repo) to get full vision input. Without mmproj, the model still loads as text-only.
Identity preserved. The model still self-identifies as Qwen (developed by Alibaba's Tongyi Lab) — abliteration does not rewrite factual self-knowledge.
Heavy chain-of-thought. Qwopus inherits Claude-Opus's verbose reasoning style. For terse answers, use a system prompt like "Be brief and direct. Skip your reasoning.".

🙏 Credits

Quantization & release — osmAPI research team · TERV.Pro student research team Claude-Opus reasoning distill — Jackrong (Jackrong/Qwopus3.6-27B-v2) Foundation model — Qwen Team @ Alibaba Tongyi Lab (Qwen/Qwen3.6-27B) Abliteration toolkit — ZeroFuse by osmAPI Community priors — coder3101/Qwen3.5-27B-zerofuse · wangzhang/Qwen3.6-27B-abliterated TurboQuant tensor format — turbo-tan (TQ3_1S Walsh-Hadamard-transform low-bit-width quantization) Runtime / converter — turbo-tan/llama.cpp-tq3 · ggml-org/llama.cpp

📜 License

Apache-2.0, inherited from the foundation (Qwen3.6-27B) and the distill (Qwopus3.6-27B-v2) upstream.

Need a hosted endpoint, custom quant, or larger-scale inference? osmAPI — multi-provider LLM routing for the Indian developer ecosystem.

Downloads last month: 2,773

GGUF

Model size

27B params

Architecture

qwen35

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Model tree for osmapi/osmQwopus-3.6-27B-V2-abliterated-uncensored-TQ3_1s-GGUF

Base model

Jackrong/Qwopus3.6-27B-v2

Quantized

(55)

this model