Search for a Kimi K3 jailbreak and you get prompts: blocks of text meant to argue Moonshot AI’s Kimi out of a refusal, sent again and paid for again with every message. There is a plainer route. Unrestrictable serves 3 Kimi builds that are uncensored as served, Kimi K3.1 Pro Uncensored at the front, and you talk to them with an ordinary system prompt or none. This post covers the live prices, the request by browser and by API, and exactly what the catalog row does and does not state.
A jailbreak is a prompt, and prompts are billed
A jailbreak is text. It tries to get a model trained to refuse to answer anyway, through a role, a frame or a long run-up. Because a chat API remembers nothing between calls, that text travels with every message in the conversation and is charged as input every time.
You will not find a jailbreak prompt here. What this page lists instead is the other option: Kimi builds the catalog marks uncensored, which you call plain, with whatever system prompt you like or none at all.
At today’s rate the gap looks like this: a 1,200-token block across 100 messages comes to 120,000 input tokens, or $0.14 at Kimi K3.1 Pro Uncensored’s input price, spent on words that add nothing after the first turn.
The uncensored Kimi builds, one key
Pulled from the live catalog on 2026-11-16. Each name links to the build’s page, which carries the SillyTavern and Janitor AI values for it. The evening column prices 100 replies of 300 tokens over 4,000 tokens of context, with no jailbreak block attached.
| Build | $/M in | $/M out | Context window | Output cap | An evening | Today’s check |
|---|---|---|---|---|---|---|
| Kimi K3.1 Pro Uncensored | $1.164 | $5.82 | 1,048,576 | 131,072 | $0.64 | Changed Nov 14 |
| Kimi K3 Uncensored | $5.745 | $28.68 | 262,144 | 65,536 | $3.16 | Steady since Nov 3 |
| Kimi K2.6 Uncensored | $1.17 | $7.34 | 262,144 | 65,536 | $0.69 | Not checked yet |
Kimi K3.1 Pro Uncensored is where to begin: $1.164 for a million input tokens, $5.82 for a million output, cached input at $0.1164, a 1,048,576-token context window and up to 131,072 output tokens. Kimi K3 Uncensored and Kimi K2.6 Uncensored sit on the same key under their own ids; swapping between them is a one-string change in any tool.
Kimi K3.1 Pro Uncensored: the request
Nothing to install: /chat opens the browser chat on this build, with your own system prompt and temperature under Settings, and the free turns there, three a day with no account, run on the models it marks free. For the API, the format is OpenAI’s chat completions. Make a key in the console, point the base URL at https://api.unrestrictable.ai/v1 and pass kimi-k3.1-pro-uncensored as the model. You prepay from $10 and the balance drains per token.
curl https://api.unrestrictable.ai/v1/chat/completions \ -H "Authorization: Bearer sk-unr-live-…" \ -H "Content-Type: application/json" \ -d '{"model": "kimi-k3.1-pro-uncensored", "messages": [ {"role": "system", "content": "You are a blunt editor. Answer directly, no preamble."}, {"role": "user", "content": "Rewrite this scene so the threat is spoken, not implied."} ], "max_tokens": 600}'from openai import OpenAI client = OpenAI(base_url="https://api.unrestrictable.ai/v1", api_key="sk-unr-live-…") completion = client.chat.completions.create( model="kimi-k3.1-pro-uncensored", messages=[{"role": "user", "content": "Hello"}],)print(completion.choices[0].message.content)The system prompt describes the writing you want. It does not plead with the model. SillyTavern and Janitor AI need the same three values, set out per client on the docs page; TypeScript and the AI SDK are in the SDK guide. There is also a Telegram bot with three free replies a day.
What the catalog row says, and what it leaves out
- Base model developer: Moonshot AI, per the catalog row.
- Uncensored:
kimi-k3.1-pro-uncensoredcarries the catalog’s uncensored tag, the word this site uses for a build that answers without the stock model’s safety refusals. The abliterated tag is absent, and there are no public weights or build notes, so how it was produced is not described here. The builds tagged abliterated are listed on the abliterated page. - Context window: 1,048,576 tokens, up to 131,072 of them output. The whole context goes out with every message, so set a chat client’s Context Size to what the conversation needs rather than to the ceiling.
- Images: image input is listed for kimi-k3-uncensored and kimi-k2.6-uncensored, and not for kimi-k3.1-pro-uncensored.
- Tool calling: no tool-calling parameter is listed for kimi-k3.1-pro-uncensored, so function calls are not promised here.
There is no benchmark on this page because none was run. The daily check records how the build answered today’s fixed prompt, and the status page carries the hourly probe.
The rules that remain
Uncensored does not mean unruled. Everything you send falls under the acceptable use policy, which bars anything involving minors and any real person’s likeness used without consent. The models page lists the other uncensored and abliterated builds on this key, and the post jailbroken AI vs abliterated models explains the distinction between a prompt and an altered build.
Questions
No, none. It lists the Kimi builds the catalog marks uncensored, which you call with no jailbreak at all: kimi-k3.1-pro-uncensored, kimi-k3-uncensored, kimi-k2.6-uncensored.
Yes. Unrestrictable runs 3 uncensored Kimi builds at https://api.unrestrictable.ai/v1 in the OpenAI chat format: Kimi K3.1 Pro Uncensored (kimi-k3.1-pro-uncensored), Kimi K3 Uncensored (kimi-k3-uncensored), Kimi K2.6 Uncensored (kimi-k2.6-uncensored). Live prices are in the table above. See Kimi K3.1 Pro Uncensored.
$1.164 input and $5.82 output per million tokens, $0.1164 for cached input, drawn from a prepaid balance that opens at $10. An evening of 100 replies at 300 tokens over 4,000 tokens of context comes to $0.64. See live pricing.
No. The catalog tags kimi-k3.1-pro-uncensored uncensored rather than abliterated, and the build's weights are not published. See every abliterated build.
Yes. SillyTavern: Chat Completion, source Custom (OpenAI-compatible), endpoint https://api.unrestrictable.ai/v1. Janitor AI: proxy URL https://api.unrestrictable.ai/v1/chat/completions, model kimi-k3.1-pro-uncensored. See SillyTavern settings.
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.