Pricing
One monthly price for the platform and the models.
Every plan includes hosted runs and standard tokens you can spend on any supported model, through one mAIvn key. Start with the Free trial and move up when your work does.
mAIvn is in invite-only beta. Beta accounts aren't charged; paid plans, top-ups and automatic reloads open at general availability.
Plans
Free trial
Build and run your first agent.
$0once
- Standard tokens
- 1,000,000 once
- Runs
- 50 once
- Concurrent jobs
- 1
Starter
Build your first workflow.
$39/ month
- Standard tokens
- 5,000,000 / month
- Runs
- 1,000 / month
- Concurrent jobs
- 2
Professional
Grow what you have built.
$149/ month
- Standard tokens
- 10,000,000 / month
- Runs
- 3,000 / month
- Concurrent jobs
- 5
Business
Cover more of your team's work.
$499/ month
- Standard tokens
- 50,000,000 / month
- Runs
- 15,000 / month
- Concurrent jobs
- 15
Enterprise
A contract sized to your usage, with terms we agree together.
Prices in USD. Paid plans bill monthly; taxes extra.
How usage works
Plans are priced in standard tokens. One million standard tokens equals $2 of model usage at the provider's list price. A model's input, output and cache tokens count in proportion to its list price, with no markup and no tier minimums.
So one allowance works across every model: lower-cost models use fewer standard tokens, and more capable models use more. Starter's 5 million standard tokens cover $10 of model usage; Professional's 10 million cover $20; Business's 50 million cover $100.
How a list price converts
- $0.10
- 50,000
- $1
- 500,000
- $2
- 1,000,000
- $10
- 5,000,000
- $25
- 12,500,000
Output usually costs more than input, and cache reads much less. Per-model rates are listed in the developer portal.
When you reach a limit
New runs pause when your standard tokens or runs run out. A run that's already going finishes, and anything it uses past your limit is taken from the next tokens you add, or from next month's allowance. You choose how to keep going.
The order your tokens are used in
Your plan's allowance
Used first. It resets at the start of each billing month, and unused allowance doesn't roll over.
Tokens you buy
Used after your allowance runs out. They never expire.
Top-ups are not available on the Free trial.
Questions
The unit your plan's allowance is counted in. One million standard tokens equals $2 of model usage at provider list price. A model's tokens count in proportion to its list price, for input, output and cache tokens alike, with no markup: a model priced at $10 per million tokens uses 5 million standard tokens for every million of its tokens. Starter's 5 million standard tokens cover $10 of model usage.
New runs pause. A run that's already going finishes, and anything it uses past your limit is taken from the next tokens you add, or from next month's allowance; it's never billed separately. To keep going now, upgrade to a bigger plan, add a top-up, or turn on auto-reload where it is available.
Your plan's monthly allowance resets each billing month and does not roll over. Tokens you buy in a top-up never expire; your monthly allowance is used first, then any tokens you have bought.
Every new account gets 1,000,000 standard tokens and 50 runs, once, to build and run a real agent. Trial tokens don't expire. Top-ups aren't available on the trial; move to a paid plan to keep going.
The pricing section of the developer portal lists the rates for more models, and every charge records the rate it was billed at.
Start with the Free trial
1,000,000 standard tokens and 50 runs to build a real agent.