Token pricing
Uncensored inference pricing breaks down into clear per-token rates for input and output, with prepaid credit that never expires and crypto-only top-ups. You pay only for what you process, and errors or refusals are free.
No subscription. Prepaid credit never expires.
- $0.25per 1M input tokens
- $1.00Output tokens / 1M
- 64,000token context
- $0.50trial credit
- 300requests per minute
Worked examples
Code completion in a developer IDE
A typical request sends 2,000 input tokens and receives 500 output tokens. Input cost: 2,000 / 1,000,000 * $0.25 = $0.0005. Output cost: 500 / 1,000,000 * $1.00 = $0.0005. Total: $0.001 per request.
Long-context document summarization
Processing a 32,000 token document with a 1,000 token summary. Input cost: 32,000 / 1,000,000 * $0.25 = $0.008. Output cost: 1,000 / 1,000,000 * $1.00 = $0.001. Total: $0.009 per request.
High-volume chat messages
A batch of 100 short chat turns, each with 500 input and 200 output tokens. Input cost: 100 * (500 / 1,000,000 * $0.25) = $0.0125. Output cost: 100 * (200 / 1,000,000 * $1.00) = $0.02. Total: $0.0325 for the batch.
Token Pricing: Input vs Output
We charge separate rates for input and output tokens. Input tokens cost $0.25 per million, while output tokens cost $1.00 per million. This reflects the higher compute cost of generating text. The model id you send is "uncensored". It is an open-weight model run on our servers, tuned to answer without content refusals for lawful adult use. It is not GPT, Claude, Gemini, Grok, DeepSeek, Qwen, Llama, or any other vendor's model.
Our API supports streaming via SSE, function calling, and JSON mode. Parameters like temperature, top_p, stop, seed, presence_penalty, and frequency_penalty are available. The context window is 64,000 tokens, with a max output of 16,000 tokens per request. Errors and refusals are free, so you only pay for successful completions.
Prepaid Credit & No Expiration
Prepaid credit is charged by real token usage. Paid credit never expires, so you can top up once and use it over time. There is no subscription and no monthly fee. You can start with a trial credit of $0.50, valid for 7 days, with no card needed. Every new account gets this trial credit, one per person.
A prepaid balance keeps spend predictable because usage can never exceed what was loaded. You can monitor your usage through the dashboard. If you make a mistake, such as a double charge, it is fixed through the Support page. Credit is not refunded, but it remains available for future requests.
Crypto Top-Up: USDT (TRC20) & USDC (Base)
Top-ups are crypto only: USDT (TRC20) or USDC (Base). You can top up any whole amount from $10 to $500. No cards, no PayPal, no bank transfer. This keeps transactions fast and global. You sign up with "Continue with Google" or email and password on the "Get API key" page. The key is shown immediately. No phone number is required.
Bonus credit is added on top-ups: +5% bonus credit from $50, +10% from $100. This bonus is added to your balance and can be used for requests. Your API key works with the official OpenAI SDKs and any OpenAI-compatible client by changing base_url to https://api.kimiapi.top/v1. Privacy is respected: an account needs only an email, and prompts are not used for training.
Questions and answers
What is the context window and max output?
The context window is 64,000 tokens, including prompt and completion together. The max output is 16,000 tokens per request, or 2,048 if max_tokens is not set. This allows for long documents and complex reasoning tasks.
How do I top up my credit?
Top-ups are crypto only: USDT (TRC20) or USDC (Base). Any whole amount from $10 to $500 is accepted. Bonus credit is added: +5% from $50, +10% from $100. No cards, PayPal, or bank transfers are supported.
Is the trial credit refundable?
No, trial credit is not refundable. It is valid for 7 days and is one per person. No card is needed to get the $0.50 trial credit. Mistakes like double charges are fixed through the Support page.
What content is blocked?
A hard content limit always applies: no sexual content involving minors. Requests of that kind are refused. The model does not refuse lawful adult, fictional, security-research, or controversial topics, making it suitable for diverse use cases.
Your key is one form away
Create an account, copy the key, change the base URL. That is the whole setup.
