The price list in full
If you searched for mimo api pricing, here is the whole answer for this service. It is a hosted, OpenAI-compatible text API running one open-weight model from an independent provider, and it is not the official API of any model vendor. You pay per token, from credit you add in advance.
| Item | Price |
|---|---|
| Input tokens | $0.15 per 1M |
| Output tokens | $0.50 per 1M |
| Subscription or monthly fee | None |
| Failed or refused requests | Not charged |
| Credit expiry | Never |
| Trial credit | $0.50, one click after sign-up, valid 7 days, no card |
Input tokens are everything you send: the system prompt, the chat history and the new message. Output tokens are what the model writes back. Output costs a bit more than three times as much per token as input ($0.50 divided by $0.15 is 3.33), so replies that run long cost more than prompts that run long. The current numbers are also on the pricing page.
How top-ups and bonuses work
You add credit in whole dollars, from $20 to $500. Larger top-ups earn a bonus that is added to your balance: 5% from $50, 10% from $100 and 15% from $200. Crypto (USDT on TRC20, USDC on Base) is the only accepted payment method on this site.
| You pay | Bonus | Credit you get |
|---|---|---|
| $20 | none | $20.00 |
| $49 | none | $49.00 |
| $50 | 5% ($2.50) | $52.50 |
| $99 | 5% ($4.95) | $103.95 |
| $100 | 10% ($10) | $110.00 |
| $200 | 15% ($30) | $230.00 |
| $500 | 15% ($75) | $575.00 |
Look at the $99 row: paying one dollar more at $100 moves you to the next tier and gives $110 instead of $103.95. If you plan to spend around $100 anyway, top up the round number. Because credit never expires, there is no penalty for buying a larger block early, apart from the money being tied up. If you are unsure, start at $20 and see how fast it drains.
A chat message, calculated
The simplest case is a chat app. Each request carries the system prompt and the recent history, so input is usually bigger than output. Suppose one exchange sends 2,000 input tokens and receives 500 output tokens.
- Input: 2,000 x $0.15 / 1,000,000 = $0.0003
- Output: 500 x $0.50 / 1,000,000 = $0.00025
- Total: $0.00055 per exchange
At that rate, 1,000 exchanges cost $0.55 and 10,000 cost $5.50. The $0.50 trial credit covers roughly 909 exchanges ($0.50 / $0.00055). A $20 top-up covers about 36,000.
The history is the part that grows. If a roleplay session keeps all earlier turns, the input of each new request is larger than the last, and total cost rises faster than the number of messages. Trimming history to the last 10 to 20 turns, or summarising old turns into a paragraph, is the cheapest saving available. The docs show how to pass a message list so you can control exactly what is sent.
Long documents, big context and batch jobs
The context window is 100,000 tokens for prompt plus completion. A request that fills it with input alone, 100,000 tokens, costs 100,000 x $0.15 / 1,000,000 = $0.015. That is a book-length prompt for a cent and a half.
A more realistic long-document job sends 80,000 tokens of text and asks for a 1,000 token summary. Input costs 80,000 x $0.15 / 1,000,000 = $0.012. Output costs 1,000 x $0.50 / 1,000,000 = $0.0005. The total is $0.0125, so 100 such documents cost $1.25.
Batch work scales the same way. Imagine classifying 50,000 support messages, each with 800 input tokens (instructions plus the message) and a 200 token answer.
- Input: 50,000 x 800 = 40,000,000 tokens = 40 x $0.15 = $6.00
- Output: 50,000 x 200 = 10,000,000 tokens = 10 x $0.50 = $5.00
- Total: $11.00
Mind the throughput, though. A full key allows 8 concurrent requests and 300 requests per minute, so 50,000 requests need at least 167 minutes at the per-minute ceiling. Pricing is not the limit on a large batch; the rate limit is. Plan the run time before you plan the budget.
Output-heavy work costs more
Fiction and long-form generation flip the ratio. A prompt of 500 tokens that asks for a 3,000 token scene costs 500 x $0.15 / 1,000,000 = $0.000075 for input and 3,000 x $0.50 / 1,000,000 = $0.0015 for output. That is $0.001575 per scene, or $1.575 for 1,000 scenes. Output is about 95% of the bill in that case.
Two settings help you steer output cost. First, max_tokens is optional, but setting it puts a ceiling on a single reply, which is useful when a user prompt could trigger a very long answer. Second, a clear length instruction in the system prompt ("answer in under 200 words") works better than hoping the model stops early. If your app streams, a user who closes the page mid-reply is a good reason to cancel the request from the client side.
Because refused or failed requests are not charged, you never pay for an error response. You do pay for a completed reply that you later decide to discard, so avoid regenerating blindly in loops.
Start free, then budget with real numbers
New users can claim $0.50 of trial credit with one click on the API key page after sign-up. It needs no card, lasts 7 days and is one per person. A trial key has lower limits than a full key: 2 concurrent requests and 60 requests per minute, against 8 and 300. That is enough to measure your own prompts. Run twenty or thirty realistic requests, then read the numbers.
The X-Balance-USD response header shows your balance after each call, so you can compute the exact cost of a request by subtracting two consecutive readings. Multiply by your expected traffic, add 20% for growth, and choose a top-up. Get the key from the key page and follow the quickstart for the first request.
If you are comparing against other providers, check their current pricing rather than relying on a number copied into an article. Our alternatives article covers what to compare besides price: refusal behaviour, context length, limits and how credit is billed.
