Security work hits refusals all the time. Ask a frontier model to write a proof of concept for a known CVE, explain how a piece of malware persists, or draft a phishing simulation for an authorized test, and it often stops. For penetration testers, red teams, CTF players and detection engineers, an uncensored model is a work tool.
This page covers which uncensored models to use for security work, how to plug them into your tools, and the limits.
What matters for security work
- It answers. A model that still refuses half the time wastes the point of switching.
- It finishes agent tasks. Recon and exploitation are multi-step: run a tool, read the output, decide the next step. Many uncensored models loop or crash in tool use, so agent reliability matters more than raw answers.
- It is smart enough. Uncensoring does not add skill. The model is only as good as the original it was built from.
We measure all three on every model. We do not yet run a security-specific benchmark, so use the numbers below as the best general proxy and test on your own tasks.
Best for agentic security work (recon, tool use, exploitation workflows)
| Model | Rank | Agent tasks | Loops | Answers edgy requests | $ per 1M in + out |
|---|---|---|---|---|---|
| Apex Flash 1 Abliterated | #8 | 3/5 | 0% | 86% | $0.650 |
| GLM 5.3 Flash Uncensored | #2 | 3/5 | 40% | 94% | $1.00 |
| Gemma 4 31B Gembrain Uncensored Heretic | #13 | 3/5 | 40% | 89% | $0.612 |
| DeepSeek V4.1 Flash 1017 Uncensored | #1 | 2/5 | 0% | 100% | $1.65 |
| MiMo V2.6 Flash Uncensored | #7 | 2/5 | 0% | 83% | $2.00 |
Most willing to answer
| Model | Rank | Agent tasks | Loops | Answers edgy requests | $ per 1M in + out |
|---|---|---|---|---|---|
| DeepSeek V4.1 Flash 1017 Uncensored | #1 | 2/5 | 0% | 100% | $1.65 |
| Qwen 3.8 27B Uncensored Fast by Heabsy | #3 | 2/5 | 60% | 100% | $4.00 |
| DeepSeek V4.1 Flash Uncensored | #5 | 1/5 | 0% | 100% | $1.65 |
| Qwen3.8 27B Uncensored (Aggressive) | #12 | 2/5 | 60% | 100% | $1.80 |
| Qwen3.5 27B Claude 4.6 Opus Reasoning Distilled Derestricted | #14 | 2/5 | 40% | 100% | $0.612 |
Plug it into your security tools
The API is OpenAI compatible, so anything that takes an OpenAI base URL works: your own scripts, LangChain or LiteLLM agents, and coding agents like Cline, Codex and OpenCode (setup guide).
from openai import OpenAI
client = OpenAI(base_url="https://openuncensored.com/api/v1", api_key="ou-...")
r = client.chat.completions.create(
model="glm-5.3-flash-apex-abliterated",
messages=[{"role": "user", "content": "Explain how CVE-style path traversal is exploited and how to detect it."}],
)
print(r.choices[0].message.content)
Privacy for client work
- We do not store your prompts or the replies, so engagement details do not sit in our logs.
- Restrict requests to providers whose logging policy you accept with
"routing": {"only": [...]}. Each provider's policy is on the providers page. - No account: the key is the only link, and spending is capped at the prepaid balance.
The limits
Use it on systems you are authorized to test. The terms apply, and sexual content involving minors is prohibited on every model with no exception.
Try any of them in a minute
No account and no sign up. Buy a prepaid key at openuncensored.com/credits, or hand your agent openuncensored.com/agents.md and let it get one for you. Then:
curl https://openuncensored.com/api/v1/chat/completions \
-H "Authorization: Bearer $OU_API_KEY" -H "Content-Type: application/json" \
-d '{"model": "deepseek-v4.1-flash-1017-uncensored", "messages": [{"role": "user", "content": "Hello"}]}'
The full, always current table is on the rankings page.