People search for an uncensored ChatGPT after a refusal. OpenAI does not offer one, so the working answer is another model: 158 open-weight builds with the refusal removed, free to try in the chat and callable from the code you already wrote for ChatGPT.
Use an uncensored ChatGPT alternative now
Three ways in, from no account to your own app.
- No sign-up: the chat on the home page gives a few free turns a day on 95 models, with replies up to 1,000 tokens.
- A free account: 5 chat messages a day in the console on deepseek-v4.1-flash-1017-uncensored, with no card.
- A key: all 158 chat models per token, prepaid from $10, in the OpenAI SDK or any OpenAI-compatible client. The docs have the base URL, the key and the model id.
ChatGPT and Unrestrictable side by side
The ChatGPT column is from its pricing page, read 10 November 2026. Ours is read from the live catalog when this page renders.
| What you get | ChatGPT | Unrestrictable |
|---|---|---|
| Refusals | Usage must comply with OpenAI’s policies | Every listed model has the refusal removed or tuned out; chat requests are not filtered for topic |
| Free chat | Unlimited text chats on the Free plan, subject to abuse guardrails | A few free turns a day with no sign-up; 5 console messages a day on a free account |
| Models | OpenAI’s own models | 158 open-weight chat models: abliterated, uncensored, roleplay and security builds |
| Paying | Monthly plans: Go $8, Plus $20, Pro from $100 | Prepaid from $10, priced per token, no subscription; credit lasts 12 months |
| From your own code | The OpenAI chat format | The same format: change the base URL, the key and the model id |
Why ChatGPT refuses and these models answer
ChatGPT is a service, and its rules are part of it: the plans page says usage must comply with OpenAI’s policies, and no setting turns that off. A jailbreak prompt is a workaround you paste and keep repairing.
The models here are a different kind of thing. Each is an open-weight model whose refusal behaviour has been removed from the weights or tuned out, so there is nothing to paste. Unfiltered ChatGPT: what exists instead goes through why, and abliterated vs uncensored explains the two ways a build gets there.
Builds to start with, priced live
Per million tokens, read from the catalog at render. The leaderboard ranks every model by today’s check, and the models page lists all 169. The daily real-model check read 146 of 169 steady today.
| Wire formats | |||||
|---|---|---|---|---|---|
| GLM-5.3 Uncensored | Uncensored | $0.282 | $0.846 | — | |
| DeepSeek V4.1 Flash Abliterated | Abliterated | $0.8488 | $1.6975 | — | |
| GLM-5.3 Flash Uncensored | Uncensored | $0.1053 | $0.3685 | — | |
| Qwen3.8 27B Uncensored | Uncensored | $1.2125 | $3.6375 | 33K | |
| Llama 3.3 70B Abliterated | Abliterated | $2.80 | $2.80 | 33K |
Keep the OpenAI SDK, change the base URL
Code written for ChatGPT’s API runs here with three values changed. Before:
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])response = client.chat.completions.create(model="<an OpenAI model>", messages=[...])After:
client = OpenAI(base_url="https://api.unrestrictable.ai/v1", api_key=os.environ["ULLM_API_KEY"])response = client.chat.completions.create(model="glm-5.3-uncensored", messages=[...])Create the key in the console and top up from $10. The docs page has the same three values for SillyTavern, Janitor AI and the SDKs.
Where ChatGPT is the better pick
- Your questions never meet a refusal. ChatGPT’s free plan has unlimited text chats and you do not need this site.
- You want OpenAI’s own models, voice chats, deep research or memory. Its plans list all four.
- You want an app from the App Store or Google Play. ChatGPT lists web, iOS and Android.
Only want to chat in a browser?
The API is built around the key. The chat runs the same models with images and video and no key to manage.
Questions
Not from OpenAI. Its plans page says usage must comply with its policies (read 10 November 2026). An uncensored ChatGPT alternative is a different service running open-weight models with the refusal removed, which is what this catalog is.
A jailbreak prompt is a workaround that has to be pasted and maintained. An abliterated model has the refusal direction removed from its weights, so there is nothing to paste and nothing to maintain.
To try, yes. A few free turns a day on 95 models with no sign-up, and 5 console chat messages a day on a free account. The API is prepaid from $10.
No. Unrestrictable is not ChatGPT and is not an OpenAI product. It serves open-weight models through the same API format.
Yes. Set the base URL and the key and pick a model id from the models page. Chat completions, streaming and tool calls use the OpenAI format.
The acceptable use policy lists what is never allowed, and image and video prompts involving minors are refused before they are priced. See the acceptable use policy.
Every model above is on the uncensored model list with its live price per million tokens. Per-token pricing sets the whole catalog side by side, and the family directory groups the builds by base model.