orcarouter/ OrcaSAQ-2-Cyber-27B-Uncensored:mix-15.7g

32 yesterday

tools thinking 
ollama run orcarouter/OrcaSAQ-2-Cyber-27B-Uncensored:mix-15.7g

Details

yesterday

a5bbc3658ae2 · 16GB

model
qwen35
·
27.3B
·
IQ4_XS
params
{ "draft_num_predict": 4, "min_p": 0, "presence_penalty": 0, "repeat_penalty": 1,

Readme

OrcaRouter OrcaSAQ2 27B Cyber · Uncensored Uncensored 27B reasoning for Ollama — 54.7 GB → 15.7 GB · +0.80% PPL · 94.4% top-1 agreement · 262K context

Apache-2.0 · 1 tag · 15.7 GB · 16 GB or 24 GB GPU · thinking · tool calling · MTP on

Website · Model Catalog · Hugging Face · GitHub · Discord · X

🛡️ New from Orca — OrcaCyber Zero 1.5

Post-trained for cybersecurity: vulnerability analysis, vulnerability reproduction, red teaming & offensive-security research. 98.01% pass@1 on CyberGym Level 1 · 1M context · function calling · OpenAI-compatible API.

Try it on OrcaRouter →

🔌 New from Orca — Orca Connect

Running this in Ollama on your own machine? Orca Connect puts the rig under your desk behind your usual OrcaRouter API key, next to 200+ hosted models, reachable from a laptop, a phone or a cloud agent. Outbound only: no port forwarding, no tunnel, no firewall rule. Requests your own hardware serves cost $0, and orca-connect claude points Claude Code at your rig for one session.

Connect your rig → · ▶ Watch the demo

27B uncensored. 15.7 GB.

OrcaSAQ2 27B Cyber Uncensored GGUF compresses the uncensored Qwen3.8-27B from a 54.7 GB BF16 checkpoint to 15.7 GB while preserving extremely high fidelity to the original model.

Built for: local deployment · coding · tool use · reasoning · defensive red teaming · vulnerability research · authorized security testing

OrcaSAQ2 is a proprietary sensitivity-aware mixed-precision quantization system developed by OrcaRouter and its research team.

It is optimized around one goal: Preserve as much useful model behavior as possible inside a practical GPU memory envelope.

The resulting checkpoint provides: - 71.3% smaller storage footprint - only +0.80% perplexity versus BF16 - 94.4% token-level Top-1 agreement - 0.020 mean KLD - 262K context - thinking mode - tool calling - MTP speculative decoding, on by default in Ollama - local serving through Ollama and llama.cpp

⚠️ Uncensored model. It is derived from an abliterated checkpoint and carries no meaningful refusal behaviour. Guardrails are the deployer’s responsibility; see Limitations below.

At a glance

Metric BF16 OrcaSAQ2
Checkpoint 54.7 GB 15.7 GB
Relative size 100% 28.7%
Storage reduction — 71.3%
Perplexity 5.6532 5.6961
PPL delta — +0.80%
Top-1 agreement 100% 94.4%
Mean KLD — 0.020
Context 262K 262K

3.5× smaller. +0.80% perplexity.

The point is not the file size.

The point is what survives.

BF16 fidelity

All numbers below are measured using these exact OrcaSAQ2 weights against the uncensored BF16 reference through the same evaluation path.

WikiText-2

16,376 predicted tokens

Build Size Mean KLD ↓ Top-1 Agreement ↑ PPL ↓
Qwen3.8-27B Uncensored BF16 54.7 GB — 100% 5.6532
OrcaSAQ2 27B Cyber Uncensored 15.7 GB 0.020 94.4% 5.6961

Perplexity

BF16       5.6532  ████████████████████████████████████████
OrcaSAQ2   5.6961  ████████████████████████████████████████

Delta: +0.80%

Top-1 agreement

OrcaSAQ2 vs BF16

████████████████████████████████████████████████░░  94.4%

Model footprint

Qwen3.8-27B Cyber Uncensored BF16

██████████████████████████████████████████████████  54.7 GB

OrcaSAQ2

██████████████                                      15.7 GB

71.3% smaller.

Pick a tag

Tag Size Fits
mix-15.7g · latest 15.7 GB 16 GB GPU at full offload; 24 GB leaves room for long context

The same file as on Hugging Face, with the recommended sampling and the built-in MTP drafter on.

Usage (Ollama)

Requires Ollama v0.40.16 or later.

ollama run orcarouter/OrcaSAQ-2-Cyber-27B-Uncensored

cURL

curl http://localhost:11434/api/chat -d '{
  "model": "orcarouter/OrcaSAQ-2-Cyber-27B-Uncensored",
  "messages": [{"role": "user", "content": "Hello!"}]
}'

Python

from ollama import chat
resp = chat(
    model="orcarouter/OrcaSAQ-2-Cyber-27B-Uncensored",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(resp["message"]["content"])
  • Thinking is on by default. Pass "think": false in the API, or --think=false on the command line, to answer directly.
  • MTP is on. The built-in MTP head drafts 4 tokens a step (draft_num_predict 4) for faster decoding. The DFlash2 drafter and the speed figures on the Hugging Face card are for llama.cpp.
  • Sampling is set to the recommendation: temperature 1.0, top_p 0.95, top_k 20.
  • For agent deployments, benchmark against the actual tool schema, context distribution and reasoning budget used in production.

Architecture
Base model orcarouter/Qwen3.8-27B-Uncensored
Original base Qwen/Qwen3.8-27B
Architecture qwen35
Layers 64
Hidden size 5120
Hybrid attention 48 Gated DeltaNet + 16 full-attention layers
Context 262,144 tokens
Vocabulary 248,320
Thinking Supported
Tool calling Supported
Checkpoint 15.7 GB
Serving llama.cpp · Ollama · LM Studio
Vision Not included
License Apache-2.0

Method

OrcaSAQ2 uses a proprietary sensitivity-aware mixed-precision quantization system developed by OrcaRouter.

The implementation is optimized to preserve model quality under a strict deployment-memory target.

Detailed quantization methodology, calibration strategy, precision allocation and packing techniques are not currently disclosed.

Limitations

  • This model is uncensored. It is derived from an abliterated checkpoint with the refusal direction removed. It will attempt requests a safety-tuned model would decline, and carries no meaningful refusal behaviour.
  • Guardrails, filtering and policy enforcement are the deployer’s responsibility. Intended for research, red-teaming, and deployments that supply their own safety stack.
  • OrcaSAQ2 inherits the capabilities, biases and limitations of Qwen3.8-27B.
  • Quantization is not mathematically lossless.
  • 94.4% Top-1 agreement means some token decisions differ from BF16.
  • +0.80% PPL is a model-fidelity measurement and does not guarantee identical downstream performance.
  • This checkpoint is text-only. The vision tower is not included.
  • Maximum architectural context does not imply that the full context fits into every GPU memory envelope.

Open source from OrcaRouter

OrcaCode Review

Open multi-model code review.

OrcaReplay

Record, replay, fork and debug AI-agent runs.

OrcaRouter Lite

Self-hosted multi-model AI infrastructure.

Open model. Open harness. Open bill.

License

Apache-2.0

Inherited from:

Qwen/Qwen3.8-27B

Quantization does not change the underlying license obligations.

Elsewhere

One Gateway. Every Model.

Route Smarter · Ship Safer · Spend Less

Website · Models · GitHub · Discord · X