kimiapikey.comGuide
Kimi API Key Pricing: What This Alternative Costs
This page lays out exactly what a key on this site costs: two token rates, prepaid credit and no monthly fee. Every example is computed from the published rates so you can check the math yourself.
Updated

Key points
- Input tokens cost $0.15 per 1M and output tokens cost $0.50 per 1M.
- You pay from prepaid credit. Top-ups run from $20 to $500, with +5%, +10% or +15% bonuses from $50, $100 and $200.
- Failed or refused requests are not charged, and credit never expires.
- New users can claim $0.50 of trial credit with one click, with no card.
The rates in one table
People who search for kimi api key pricing usually want one thing: a number they can plug into a budget. Here it is. This site does not sell the official Kimi API. It sells access to a single open-weight model run by an independent provider, billed per token from a prepaid balance.
| Item | Price |
|---|---|
| Input tokens | $0.15 per 1M |
| Output tokens | $0.50 per 1M |
| Monthly fee | None |
| Minimum top-up | $20 |
| Maximum top-up | $500 |
| Credit expiry | Never |
| Failed or refused requests | Not charged |
Input tokens are everything you send: the system prompt, the conversation history, attached documents and the new question. Output tokens are what the model writes back. Both are counted by the model's tokenizer, and the response includes a usage block with the exact counts, so your bill never has to be estimated after the fact. The pricing page carries the same figures.
How to turn tokens into dollars
The formula is short. Cost equals input tokens divided by 1,000,000 times $0.15, plus output tokens divided by 1,000,000 times $0.50. Output is a little over three times as expensive as input per token, but most workloads send far more than they receive, so input often decides the bill.
A rough conversion helps for English text: a token is a bit under a word, so 1,000 tokens is on the order of 700 to 750 words. Code and non-English text usually use more tokens per word, so measure with the usage field once instead of trusting the rule.
Four worked examples, each computed from the rates above.
A single chat question
A prompt of 2,000 tokens and a reply of 500 tokens:
- Input: 2,000 / 1,000,000 x $0.15 = $0.0003
- Output: 500 / 1,000,000 x $0.50 = $0.00025
- Total: $0.00055
A thousand questions like that cost $0.55.
A long conversation turn
A chat that has grown to 20,000 tokens of history, with an 800-token answer:
- Input: 20,000 / 1,000,000 x $0.15 = $0.003
- Output: 800 / 1,000,000 x $0.50 = $0.0004
- Total: $0.0034
A request that fills the whole window
The context window is 100,000 tokens, prompt plus completion. A prompt of 100,000 tokens would leave no room, so take 98,000 tokens in and 2,000 out:
- Input: 98,000 / 1,000,000 x $0.15 = $0.0147
- Output: 2,000 / 1,000,000 x $0.50 = $0.001
- Total: $0.0157
Under two cents for the largest request the model can take.
A daily workload
A script that makes 200 requests a day, each with 3,000 input tokens and 700 output tokens:
- Input: 200 x 3,000 = 600,000 tokens, so $0.09
- Output: 200 x 700 = 140,000 tokens, so $0.07
- Total: $0.16 per day, or $4.80 over 30 days
What a top-up buys
You add credit in whole dollar amounts from $20 to $500. Bigger top-ups carry a bonus that is added to your balance:
| You pay | Bonus | Credit you get |
|---|---|---|
| $20 | None | $20.00 |
| $50 | +5% | $52.50 |
| $100 | +10% | $110.00 |
| $200 | +15% | $230.00 |
| $500 | +15% | $575.00 |
To see what $20 means in tokens, divide by the rate. Spent only on input, $20 buys about 133 million tokens ($20 / $0.15 per 1M). Spent only on output, it buys 40 million ($20 / $0.50 per 1M). Real use is a mix. At the 2,000-in, 500-out question from above, $20 covers about 36,000 requests.
The daily workload example costs $4.80 a month, so a single $20 top-up would last about 125 days at that pace ($20 / $0.16). Because credit never expires, there is no pressure to use it up before a deadline. The balance is also in every response: the X-Balance-USD header shows what is left after each call, so you can alert yourself in code before it reaches zero.
The free trial: what $0.50 covers
After signing up with Google, or with an email and password, the API key page shows a button for trial credit. One click adds $0.50. It needs no card, it is limited to one claim per person, and it stays valid for 7 days.
At the $0.00055 cost of the 2,000-in, 500-out question, $0.50 pays for around 900 of them ($0.50 / $0.00055 = 909). That is enough to run a real evaluation: your own prompts, your own client code, streaming and tool calling if you use them.
The trial key has lower limits than a funded key: 2 concurrent requests and 60 requests per minute. Those are fine for testing one app by hand and tight for a load test. Funded keys get 8 concurrent requests and 300 requests per minute. If you plan to test with a script, keep it to two parallel workers on the trial.
Estimate your own monthly cost in three steps
The examples above are samples, and your numbers will differ. A short method gives you a figure you can trust.
- Measure. Send 20 or so representative requests during the trial and note
prompt_tokensandcompletion_tokensfrom each usage block. Take the average of each. - Multiply. Multiply those averages by your expected requests per month. Say the averages are 1,800 input and 400 output, and you expect 50,000 requests. That is 90 million input tokens and 20 million output tokens.
- Price it. 90 x $0.15 = $13.50 for input, and 20 x $0.50 = $10.00 for output, so $23.50 a month.
Add a margin of 20 to 30 percent for retries of your own and for spikes. Then decide the top-up size. In this case a $50 top-up (which becomes $52.50 with the 5% bonus) lasts a little over two months.
Where the hidden costs usually are
There is no seat fee, no platform fee and no minimum monthly spend here, but your own code can still make a bill larger than you expect. Four things to watch:
- Resent history. A chat app sends the whole conversation each turn. A thread of 30 turns pays for turn one thirty times. Trim or summarize old messages once a thread gets long.
- Verbose system prompts. A 3,000-token system prompt on 10,000 requests adds 30 million input tokens, which is $4.50. Worth a rewrite if it is bloated.
- Unbounded output.
max_tokensis optional, and a completion can use the rest of the window. Set it when you know the answer should be short. - Retry storms. Failed and refused requests are not charged, so retries on errors are free, but a loop that retries successful calls will bill every time.
The first of these is the one that most often surprises people. A 100-turn roleplay or support thread can easily reach a context of tens of thousands of tokens, and every new message pays for all of it again.
Paying, and what is not included
Crypto (USDT on TRC20, USDC on Base) is accepted for top-ups. Prompts are not used for training.
The price covers text chat completions through the /v1/chat/completions endpoint, with streaming, tool calling and JSON mode included at the same token rates. It does not cover image, audio or video generation, embeddings or fine-tuning, because the service does not offer them. If you need those, you will want a second provider next to this one.
For a comparison with the official route, the Kimi API key explainer describes what each option is. This service is not the official Kimi API and does not claim the official vendor's prices, so check their current pricing directly if you want a side-by-side figure. When you are ready to try the rates, create your key and claim the trial. The docs show how to point an existing OpenAI client at the endpoint.
More guides
- Use a Kimi-Style API Key in Cursor: Setup Guide
- Kimi API Key Integration with OpenAI SDKs and LangChain
- Kimi API Key: How It Works and What the Alternatives Are
Frequently Asked Questions
Is the model the same as Moonshot's official Kimi?
No, we host an open-weight model that is tuned to be uncensored. It is not GPT, Claude, Gemini, Grok, DeepSeek, or any other vendor's model. It is a distinct model run on our own servers.
Do you train on my data?
No. Your prompts and completions are not used for training. We value privacy and transparency, so your data remains yours.
What are the rate limits?
You are limited to 300 requests per minute per key. The maximum request body size is 8 MB. These limits are designed to ensure stable performance for all users.
What is the content policy?
The model is uncensored regarding adult, controversial, and creative topics. However, we do block sexual content involving minors. This is a hard limit that always applies, regardless of your subscription level.
Questions and answers
Is there a monthly fee?
No. You top up prepaid credit and pay only for the tokens you use. Credit never expires.
How much does the free trial give?
New users can claim $0.50 with one click on the API key page. It needs no card, is one per person and is valid for 7 days.
Do I pay for failed requests?
No. Failed or refused requests are not charged.
What is the smallest amount I can add?
The minimum top-up is $20, and the maximum is $500. Bonuses of 5%, 10% and 15% apply from $50, $100 and $200.
Are the rates the same for streaming and tool calling?
Yes. Streaming, tool calling and JSON mode are billed at the same $0.15 and $0.50 per 1M token rates.
Your key is one form away
Create an account, copy the key, change the base URL. That is the whole setup.