Plans and pricing

Every plan comes with a monthly allowance of weighted tokens: a turn counts its prompt plus completion tokens, times the model's multiplier (GRT Nano 1×, Nano Pro 5×, Nano Max 20×). An allowance therefore buys the same dollars of list price whichever model spends it.

API keys are on every plan, Free included: the plans differ on the allowance and the rate limits, not on whether you can call the API from your own code.

These are the plans as they ship. If your deployment has been set up differently, the Plans page inside the app shows what applies to you.

Free

$0

• 1,000,000 weighted tokens a month
• GRT Nano
• 25 conversations, 1 seat
• 10,000 tokens and 20 requests a minute, 2 in parallel
• API keys
• Stops at the allowance; no overage

Pro

$20 / month, or $200 / year

• 20,000,000 weighted tokens a month
• GRT Nano and GRT Nano Pro
• Unlimited conversations, 1 seat
• 100,000 tokens and 120 requests a minute, 8 in parallel
• API keys and custom instructions
• Overage at list price, capped at $20 a month by default

Pro Plus

$50 / month, or $500 / year

• 60,000,000 weighted tokens a month
• GRT Nano, GRT Nano Pro and GRT Nano Max
• Unlimited conversations, 1 seat
• 300,000 tokens and 300 requests a minute, 16 in parallel
• API keys and custom instructions
• Overage at list price, capped at $50 a month by default

Team

$30 / seat / month, or $300 / seat / year

• 30,000,000 weighted tokens per seat a month, pooled across the organisation
• GRT Nano, GRT Nano Pro and GRT Nano Max
• Unlimited conversations and seats
• 600,000 tokens and 600 requests a minute, 32 in parallel
• API keys, custom instructions, single sign-on
• Overage at list price, capped at $30 × seats a month by default

How the allowance is spent

Model	Input / 1M	Output / 1M	Multiplier	What 1,000,000 weighted tokens buys
GRT Nano	$0.10	$0.40	1×	1,000,000 tokens of prompt + completion
GRT Nano Pro	$0.50	$2.00	5×	200,000 tokens
GRT Nano Max	$2.00	$8.00	20×	50,000 tokens

A 500-token turn on GRT Nano costs 500 weighted tokens; the same turn on GRT Nano Max costs 10,000. Usage, the allowance and what is left are on the Usage page in the app and in the /api/me/usage payload every client reads.

Overage

On a paid monthly plan, usage past the allowance is billed at the model's list price up to the organisation's overage cap. The default cap is the plan's own monthly price (per seat on Team), so a runaway key can bill at most what you already agreed to pay once; an owner can raise or lower it in Settings. Requests are refused once the cap is reached.

A yearly subscription has no overage: it gets the allowance and stops there, like Free.

Seats

Pro and Pro Plus are one seat. Team is sold per seat: the price, the allowance and the default overage cap each scale with the number of members, and the allowance is one pool the whole organisation draws on, so a five-seat team shares five seats' worth.

Changing plans

Upgrades take effect on the next request. A downgrade never deletes anything: if an organisation holds more conversations than the new plan allows, the existing ones stay readable and exportable and creating new ones is refused until you upgrade or remove some. Cancelling keeps the plan to the end of the paid period.

Invoices and enterprise terms

Organisations that do not pay by card can be billed by invoice on net terms, with a PO number. See Enterprise and invoicing.