Billing and usage
See your plan and its credit, subscribe or change plan, top up, and see what your organization has used and paid. Prices are on the pricing page.
Plans
The plan and its status are at the top of "Organization" → "Billing", and the plans side by side under "Plans".
| Plan | How it works | How you pay | Status shown |
|---|---|---|---|
| Free | Every new organization starts here: $5 of credit a month, back on the 1st (UTC). At most 1 sandbox at once and 60 minutes a run. | Free, no card | The credit left this month |
| Pro | $20 a month per organization, all of it credit, renewed each billing period. At most 3 sandboxes at once and 240 minutes a run. | Card, monthly, through Stripe | "Active" |
| Ultra | $200 a month per organization, all of it credit, renewed each billing period. At most 10 sandboxes at once and 480 minutes a run. | Card, monthly, through Stripe | "Active" |
| Prepaid | Usage is taken from a balance you top up in advance. At most 3 sandboxes at once and 240 minutes a run. | Card top-ups on the billing page, or bank transfer: we add each transfer to your balance once it arrives | "Managed by hand" |
| Enterprise | A contract, with the organization policy adjusted to you and a signed Data Processing Agreement. | Invoice | "Managed by hand" |
All amounts are in US dollars. A plan managed by hand changes when you ask us; write to admin@nice2.io.
Credit and balance
On Free, Pro and Ultra, everything metered comes out of the period's credit first: sandbox time, the judge model, platform models, the assistant and storage. On Free the period is the calendar month (UTC); on Pro and Ultra it is the billing period, which starts on the day you subscribed. The billing page, the usage page and the overview page say what is left, for example "$3.20 of this month's free credit left, resets on Nov 1, 2026". What the credit doesn't use isn't carried over.
Past the credit, usage comes out of the organization's credit balance, which holds what you topped up by card and the credit codes you redeemed. The balance never expires and stays when you change plan. Once the credit and the balance are both used up, counting what the sandboxes running now have cost so far:
- new tests on the platform executor or in a sandbox can't start, and the platform's models stop answering for runs in progress;
- a sandbox already running finishes its run (at most your plan's sandboxes at once, each for at most its longest run); what it costs past the limit comes off your next top-up;
- the credit comes back with the next period: at 00:00 UTC on the 1st on Free, on the renewal day on Pro and Ultra.
To keep going before then, top up, redeem a credit code, or move to a bigger plan. Nothing is ever billed as overage.
Subscribe to Pro or Ultra
- Open "Organization" → "Billing". Only an owner can subscribe.
- Under "Plans", click "Upgrade to Pro" or "Upgrade to Ultra". Stripe's checkout page opens; the platform never keeps card numbers.
- Pay. Back on the billing page, the status is "Active" and the period's end date is shown.
The plan's credit and limits apply from the moment you subscribe, and the first billing period starts that day. What you used on Free before that stays with that month's free credit, and your balance carries over. Where paying on the web isn't switched on yet, the page says so; write to admin@nice2.io to subscribe or to pay by invoice.
Change between Pro and Ultra
- Open "Organization" → "Billing" and click "Change plan". Only an owner can change it. Stripe's page opens on the plan change.
- Pick the other plan and confirm.
- Pro to Ultra takes effect at once. Stripe charges the difference for the rest of the period, and the period's credit goes up by the same share of the difference: halfway through a period, by $90. Ultra's limits apply at once.
- Ultra to Pro takes effect at the end of the period: until then you keep Ultra, its credit and its limits, and the next period is Pro.
To go back to Free, cancel; for Enterprise, write to admin@nice2.io.
Cancel
- Open "Organization" → "Billing" and click "Manage payment and invoices". Only an owner can cancel.
- On Stripe's page, cancel the subscription.
The subscription runs to the end of the period you paid for, and the billing page shows the day it ends; until then you keep the plan's credit and limits. Then the organization is back on Free: $5 of credit from that day to the end of the month, then $5 each month. Your balance stays. Changed your mind? Renew it on the same Stripe page before the period ends.
Top up
- Open "Organization" → "Billing". Only an owner can top up.
- Click "Top up" on the "Credit balance" card, and pick $10, $25, $50 or $100, or enter another amount: whole dollars, from $5 to $1,000.
- Pay on Stripe's checkout page. Back on the billing page, the amount is in the balance within seconds, once Stripe confirms the payment.
Each top-up comes with a paid invoice, in the "Invoices" list, and is a line in "Credits". Top-ups work on Free, Pro, Ultra and Prepaid; the period's credit is used first, then the balance. A top-up adds credit and changes no other limit. The spending cap, the plan's sandboxes at once and run length, sandbox hours, the assistant's daily limit, the per-run limits and what the AI may do by itself all stay as they are.
| What you see | What to do |
|---|---|
| "A top-up is a whole number of dollars from $5 to $1000" | Enter a whole number in that range. |
| "Top-ups by card aren't switched on yet…" | Write to admin@nice2.io to top up by bank transfer. |
| "Your plan is invoiced by usage, so there's nothing to top up…" | Enterprise and other invoiced plans don't keep a balance. Write to admin@nice2.io. |
Redeem a credit code
A credit code (PW-XXXX-XXXX-XXXX-XXXX, from us) adds credit to your organization's balance. It works on Free, Pro, Ultra and Prepaid.
- Open "Organization" → "Billing". Only an owner can redeem.
- Click "Redeem credit" on the "Credit balance" card.
- Type or paste the code (any case; dashes and spaces don't matter) and click "Redeem".
The amount goes into the balance at once, and the "Credits" list shows it. It counts exactly like a top-up: the period's credit is used first, then the balance, which doesn't expire. It never lifts your other limits.
| What you see | What to do |
|---|---|
| "That credit code isn't right…" | Check the code and try again. |
| "That credit code has expired." / "…has been revoked." / "…has been used up." | The code can't be used any more; ask whoever gave it to you. |
| "This code was already redeemed here." | Nothing to do: it's in your balance once. |
| "Too many credit code tries…" | Each organization has 10 tries an hour (20 from one network). Try again in an hour. |
| "Credit codes work on plans that keep a balance…" | Enterprise and other invoiced plans can't take a code. Write to admin@nice2.io. |
Each redemption is in the audit log ("Redeemed a credit code").
Set a monthly spending cap
On Enterprise, billed by usage, an owner can cap what the organization spends in a billing period. Free, Pro and Ultra don't take a cap: they stop once the credit and the balance are used up.
- Open "Organization" → "Billing" and find "Monthly spending cap" under "Limits".
- Enter the amount in dollars in "Monthly spending cap ($)" and click "Save cap".
- To take it off, click "Remove cap".
Everyone in the organization sees the cap and how much of it is used on the billing and usage pages; only an owner can change it, and each change is in the audit log. There is no cap until an owner sets one.
When the period's usage reaches the cap, counting what the sandboxes running now have cost so far:
- new tests on the platform executor or in a sandbox are refused, with a message that names the cap and how to raise it;
- the platform's models stop answering for runs in progress (the judge, a platform-model chat AI, an agent on the platform's model), and those runs end with a partial report;
- sandboxes already running finish their run, bounded by the plan's sandboxes at once and longest run, at their class's price.
The cap starts again with the next billing period, or as soon as an owner raises or removes it.
What is billed
| Item | How it is counted |
|---|---|
| Sandbox time | Per second from when the virtual machine is created until it is deleted, at least one minute each, at the class's price. Pre-runs and "Check once" count. A virtual machine that never reached the platform is not charged. |
| Judge model | The platform's judge (the project's model, see Models, or the model an LLM scorer names): in cloud runs and chat AI tests, in scorers that ask a judge (an SDK experiment's platform scorers, "Try it", calibration), in the Playground's scorers and in online scoring: the model bill plus 30%. |
| Platform models | Calls on a platform model through the AI Gateway and the Playground's prompt columns, a chat AI tested on a platform model, and the model calls of an agent in a repository or image project with no OPENAI_API_KEY in its secrets (your agent's model), at the same rate. |
| Assistant | The in-app assistant and the platform AI's other steps for a project (AI-written dataset rows and the issues it finds included), at the same rate. |
| Storage | Run bundles and evidence kept on the platform, per GB per day. |
| Traces uploaded | Traces from the SDK, OpenTelemetry and the Gateway, by volume. Listed on the usage page, not charged for now. |
Your own model calls (your agent's with its own key, your app's through the Gateway with x-provider-key, an SDK Eval's task, a dev AI's) go to your own provider account, and a local test runner uses no sandbox time. The prices are in unit prices. What the platform runs again because of its own problem isn't charged; see the platform's problem isn't charged.
See your usage
"Organization" → "Usage" shows one month at a time; pick another under "Month". Anyone in the organization can read it.
- The month's total, and what it is made of: "Sandbox", "Judge model", "Platform model (AI under test)", "Platform AI", "Storage" and, once traces arrive, "Traces uploaded" (their size and number of spans; not charged for now), each with the amount used and its cost.
- "Sandbox hours this month", against the limit an owner set in the organization policy, if any.
- The plan's credit left; with a spending cap, "Used $x of $y this month"; with a balance, the "Credit balance".
- "The platform's problem: not charged": what the platform ran again because of its own problem, by reason, with what it would have cost (see below). It isn't shown in a month without any.
- "By project" and "By day".
The figures are an estimate of usage to date, in US dollars as billed; the invoice is what counts.
The platform's problem isn't charged
What the platform runs again because of its own problem is recorded as usual, but not charged (since 19 October 2026):
- The platform's fill-in run after a sandbox was lost part way. None of it is charged: its sandbox time, its agent's calls to the platform model, its judging. It goes ahead even when the organization's credit, balance or monthly spending cap is used up, held only to each run's own model and judge spending limits.
- When the sandbox of a run or a pre-run is lost part way, its time after the last partial result it handed in. One lost before any attempt ran isn't charged at all, not even the minimum minute. One where attempts ran but no partial result ever came back is charged as usual: which part ran for nothing can't be told.
- Re-runs by the platform caused by the platform model failing or the platform's own test backend failing: both tries' sandbox time, and the re-run's model calls and judging. Since 20 October 2026 the platform model failing includes the model the platform put in place of the one asked for refusing a setting of the request, whatever the agent then replied; such an attempt isn't judged either, so no judging is charged for it. Since 5 November 2026 it includes the agent giving up on the platform's model after waiting 30 s or more with no answer.
- A retest ("Re-run these rows" on the experiment) of attempts that only the platform's model left untested (since 20 October 2026): none of it is charged. A retest that also takes attempts left untested for any other reason is billed as usual.
- A judge call the second judge model made because the first was rate limited (since 22 October 2026): charged at the second judge's price, never more than the first one's for the same tokens. What would be over isn't charged. The same holds when the platform judges a run again after it.
- A pre-run or check that failed only because of the platform's own part (since 22 October 2026), a run that was its own pre-run and failed so included: its model for the agent, its judge model, its tool proxy, or the test runner's own error. None of its sandbox time is charged; the pre-run page says "This sandbox time isn't charged." Its model calls are charged as they were made, as above.
Still charged: re-runs caused by an error in your own tools, data reset, database, keys or agent; a fault or an attack that never reached the agent; the agent unable to reach the model address on the sandbox itself; a job that "Lost contact" (its lease ran out); runs cancelled or stopped at a limit; a "Run the N that didn't run" you click; the first try's model calls before a re-run (charged as they were made).
The usage page's "The platform's problem: not charged" row gives what it would have cost this month, split by reason: "Filled in by the platform (a sandbox was lost)", "Re-runs by the platform", "A lost sandbox's time that left no result", "Retests of what the platform's model left untested", "The second judge's extra, used while the main judge was rate limited" and "Pre-runs and checks that failed only for the platform's own problem", with the hours where there is sandbox time. None of it is in the total, nor taken from the credit, the balance or the spending cap; its sandbox time doesn't count toward the month's sandbox hours, though it is in the "Sandbox" row's hours. On the job page of a platform fill-in run, the sandbox's cost reads "The platform's problem: not charged".
The billing page
"Organization" → "Billing" is the organization's billing dashboard. Anyone in the organization can read it; an owner subscribes, changes plan, tops up, redeems and sets the cap.
| Area | What's there |
|---|---|
| Cards at the top | The plan; the "Credit balance" with "Top up" and "Redeem credit"; what this period has cost so far; and the plan's credit still left, with the day it resets. |
| "Billing" | If tests can't start, why, and how to keep going: "Upgrade to Pro" or "Upgrade to Ultra" (Stripe's checkout), "Change plan" and "Manage payment and invoices" (Stripe's page). |
| "Usage" | The month's total, a bar for each day, the cost by type (sandboxes, judge model, platform models, assistant, storage) and the projects that cost most. Pick another month under the month list; "All usage" opens the usage page. |
| "Limits" | The limits that apply, each with how far along it is: the plan's credit, the monthly spending cap (Enterprise; an owner sets it here), the month's sandbox hours, the assistant today, and what the AI may do by itself, once any of it is switched on. |
| "Credits" | What went into the balance (top-ups, credit codes, corrections) and what it paid each period. |
| "Invoices" | Each invoice, with a link to view it and its PDF. |
| "Plans" | The plans side by side. |
| "Prices" | A sandbox hour of each class and a million tokens of each model, as billed. |
Invoices and receipts
- Pro and Ultra are invoiced by Stripe at the start of each billing period and paid by card. A change between them is prorated by Stripe, and so is the period's credit (moving from Pro to Ultra part way, say, the period has $136.98 rather than $200); the billing page says why, and the next period is whole.
- Each top-up has its own invoice, marked paid, to keep for your accounts.
- Both are under "Invoices" on the billing page, with "View" and the PDF, and on Stripe's page ("Manage payment and invoices"), where an owner also updates the card.
- Invoices are in US dollars. Ones issued before prices moved to US dollars (October 2026) are in pounds sterling and stay so.
- Enterprise is invoiced by us.
Prepaid balance
- Top up by card on the billing page (top up), or ask us for a top-up at admin@nice2.io and pay by bank transfer.
- A transfer is added to your balance once it arrives; the billing page shows it.
- Usage is taken from the balance from the day of the first top-up. Usage before that day isn't.
A credit code adds to the same balance, exactly like a top-up.
The balance counts what the sandboxes running now have cost so far. Top up before it runs out: once it is used up, new tests can't start and running sandboxes stop after the attempt in progress. With nothing topped up yet, tests can't start.
Spending limits and the estimate
- Per run. The platform's judge model and a platform-model chat AI each stop at about $6.50 per run, or at what is left of the credit and the balance, or of the spending cap, if that is less. Past the limit, the rest isn't judged or asked. None of that counts as passed, so the verdict is never a pass.
- Per run, a sandbox agent on the platform's model. Its model calls stop at about $6.50 per run, or at what is left of the credit and the balance, the spending cap or the prepaid balance if that is less. Past the limit, its calls are refused and the run stops after the attempt in progress, with a partial report. What the limit left untested is "Couldn't be tested" — the platform's brake, never the agent failing — and isn't charged: the run's report, its progress page and the round report say the one sentence "This run reached its spending limit (the agent spent $X on the platform model), so N more attempts weren't tested" — or, where the limit that stopped it was a dev AI connection's day or the organization's balance, that limit in place of what was spent ("This run reached a spending limit: the dev AI connection that started it used up today's allowance ($0.3 a day), so N more attempts weren't tested") — with the same N everywhere (those that didn't run, and any in which the platform's model had stopped answering), never "cancelled"; a run that reached it only at its last attempt says everything planned ran. Run what was left; every run has this limit, so with much left, fill it in over several runs.
- Per day. The assistant stops at $6.50 per organization per day on Free, $13 on Pro and $65 on Ultra, unless we set your organization's own limit. A top-up doesn't raise it. "Organization → Policy" has "The platform AI's daily limit": how much a day, what is used today, what happens at it and whom to ask for more, every figure as billed. With several of your organization's platform AI tasks at once and too little left today, the pages warn first.
- Per day, formal tests a dev AI starts. Each dev AI connection has its own "Spend at most $": $5 by default on a project on a local test runner, no daily limit by default on a repository or image project. A connection with no daily limit has only the organization's own credit, balance and spending cap, as a test started on the web: a test is checked against what the organization has left before it starts, and the platform's models stop where that runs out while it runs.
- Per period. The plan's credit and then the balance, the spending cap, and the sandbox hours an owner may set in the organization policy.
- The estimate before a run. When you start a test set against a chat AI, the start page shows how many questions and judge model calls to expect, the judging cost, the cost of the AI under test on a platform model, and the time, in US dollars as billed. It says so when the run would go past a per-run limit. For a repository or image project, "Start test" says "about N attempts, about M min, usually $x–$y" from what this project's own runs or checks cost an attempt (else your organization's runs whose agent used the same model), the functional tests, each security test set and the pre-run apart, and says what it is from (each test this launch runs at what its own past attempts in this project cost, so retesting other tests leaves its estimate as it was, a test that hasn't run yet at the project's average for its kind; every attempt weighs the same, so a few small retests neither pull it down nor stretch its range; a run stopped at its spending limit counts too; the models' spend is counted over the attempts that came to a verdict (at least half of those run), as one the platform couldn't judge — its model erred, the time limit — spends little; the range is at least 15% below and 20% above the usual figure, wider where these tests' past attempts differed more); when a part would go past one run's limit at its usual spend, or run longer than your organization's longest run, it says about how far it gets before it stops and what running it like that usually costs, then the ways to test it all, each with what it costs and takes: run it in N batches (the functional tests as several runs, one after another, each within its limits; the pre-run runs the first batch's tests), only the functional or only the security tests, each test once, or raise the agent model's limit for this test (with how much more than stopping at the limit; for this launch only, its pre-run included; only your organization's owner can, others are told whom to ask). The platform never tests less or switches to a cheaper model on its own to stay under a limit. Below it is the platform's AI: a part that didn't all pass gets the AI's diagnosis, usually so much a part; if the pre-run fails, having the platform fix the test connection costs so much more and takes so long (from the platform's latest), and what the project has cost so far. The spending limit is what one run may spend in all (the judge model, the agent's model and the sandbox together), given for the runs this launch starts: the pre-run, each batch of functional tests and each security test set are a run each; a part that would stop at the agent model's or the judge model's own limit names that limit, a run's at most of it. Beside "Let AI finish and check" is what it usually costs and takes, waits included, from the platform's latest on the same model started the same way (from nothing, or from a connection already there), its checks at what an attempt of this project costs and the tests the check will run, and the dearest; "Try again" and "Let the platform read the repo" say what they usually cost too (the latter with the automatic approval of the drafts it writes), and "Check once" says it for the tests it will run. Every page of a project says "This project has cost $Z so far (the platform AI $Y of it)", and an experiment's "Cost" says the project's total beside its own, and "AI diagnosis $X included" when the platform AI diagnosed its runs: the launch page's estimate counts that apart, so leave it out when you compare.
All the limits are listed under spending controls.
When tests can't start
| What you see | Why | What to do |
|---|---|---|
| "This month's free credit of $5.00 is used up…" | The Free plan's credit for this month is spent, and there is no balance. Tests on the platform executor or in a sandbox can't start. | Top up, redeem a credit code, subscribe to Pro, or wait for the date it resets. |
| "This month's free credit of $5.00 and your balance are both used up…" | The month's credit and the balance are both spent. | Top up, redeem another code, subscribe to Pro, or wait for the reset. |
| "This period's Pro credit of $20.00 and your balance are both used up…" (or Ultra's $200.00) | The period's credit and the balance are both spent. | Top up or redeem a code; on Pro you can also upgrade to Ultra. Otherwise wait for the next period. |
| "This month's spending has reached the … cap your organization's owner set…" | The period's usage reached the spending cap. | An owner raises or removes the cap, or wait for the next billing period. |
| "This month's sandbox hours are used up…" | The monthly hours an owner set in the organization policy are spent. | Wait for the 1st of next month, or have an owner raise the policy. |
| "Your prepaid balance is used up…" or "Nothing topped up yet…" | The balance is spent, or was never topped up. | Top up by card or bank transfer, or redeem a credit code. |
| "A payment has been failing for more than 7 days…" | The card payment failed and the 7 days' grace is over. | Update the payment method under "Manage payment and invoices". |
| "Your subscription was cancelled…" | A subscription to a plan billed by usage ended. (A Pro or Ultra subscription that ends puts you back on Free instead.) | Subscribe again, or write to admin@nice2.io. |
| "The platform has paused this organization…" | We paused the organization; every page says why. New tests can't start and running ones stop. | Contact us to resume. You can also ask us to pause your organization. |
Next steps
- Pricing: plans, unit prices and spending controls.
- Organization and members: owners, the policy and the audit log.
- Sandboxes and limits: what uses sandbox time.