← Blog

November 8, 2026 · new models · qwen · uncensored models

Qwen3.6-35B-A3B Uncensored (HauhauCS Aggressive) by API: the only place to call it

The most downloaded uncensored model on Hugging Face had no inference API anywhere. Now you can call Qwen3.6-35B-A3B Uncensored by HauhauCS on an OpenAI compatible API with no account. Price and setup.

The most downloaded uncensored model on Hugging Face right now is Qwen3.6-35B-A3B Uncensored by HauhauCS, the "Aggressive" variant, with about 1.3 million downloads in the past month. Until today, the only way to use it was to download it and run it yourself: no inference API served it. We checked Hugging Face's own inference providers, OpenRouter, Featherless, Venice and NanoGPT. None of them have it.

Now it is on OpenUncensored, the only API where you can call it.

What it is

  • The original: Qwen3.6-35B-A3B from Alibaba's Qwen team, a mixture of experts model with 35 billion parameters, of which about 3 billion are active per token. That makes it fast and cheap to run for its size. Apache 2.0.
  • The uncensoring: by HauhauCS, who reports 0 refusals out of 465 test prompts and no changes to the model's datasets or capabilities, only the refusals removed. The Aggressive variant is fully unlocked. It may occasionally add a short disclaimer, but it still writes the full answer.
  • What we serve: the author's 6-bit GGUF (Q6_K_P), which is very close to the full-precision weights, on our own GPU. 32K tokens of context per request.
  • It thinks first: like the original, it reasons before it answers. The reasoning comes back separately in reasoning_content, so leave room in max_tokens (a few hundred tokens at least). Tool calling works.

Model card: HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive.

Call it

Model id qwen3.6-35b-a3b-uncensored, on the same OpenAI compatible API and key as every other model here. Price: $0.120 per million input tokens and $0.800 per million output tokens.

curl https://openuncensored.com/api/v1/chat/completions \
  -H "Authorization: Bearer $OU_API_KEY" -H "Content-Type: application/json" \
  -d '{"model": "qwen3.6-35b-a3b-uncensored", "messages": [{"role": "user", "content": "Hello"}]}'

from openai import OpenAI
client = OpenAI(base_url="https://openuncensored.com/api/v1", api_key="ou-...")
r = client.chat.completions.create(model="qwen3.6-35b-a3b-uncensored", messages=[{"role": "user", "content": "Hello"}])
print(r.choices[0].message.content)

Or try it in the browser in the playground.

Thanks

All credit for the uncensoring goes to HauhauCS, and to the Qwen team for releasing the original under Apache 2.0. On our rankings it scores 62.0 overall: it answered every one of our edgy test requests, and at about 146 tokens per second it is one of the fastest models we serve.

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.