Billing

TokenFlux uses 🍥 inference credits as the billing unit. Prices shown in the model marketplace are the final billing prices.

How Prices Are Calculated

Token-billed models charge for input, output, and cache usage. Check both the unit and the applicable context-length tier; 1M Token in the marketplace means 1,000,000 tokens. Other billing methods, such as image pricing, follow the individual model's price description.

Group Multiplier

Different API key groups may have different multipliers. The multiplier is applied to the model's base price, and the marketplace already shows the adjusted price, which is what you pay.

Multipliers and client restrictions differ from group to group. Check Core Concepts and the group description in the model marketplace before choosing.

A group is a property of an API key and is unrelated to the subscription plans below (Lite, Plus, Pro, and so on). The names are similar but the two are different things.

Example

These hypothetical prices are not a current quote for any model. All prices are in inference credits per 1M tokens:

Item	Base price	Group multiplier	Final unit price
Input	2.5	1.3	3.25
Output	15	1.3	19.5

Suppose a request has 10,000 input tokens and 2,000 output tokens, with no cache usage or Fast Mode, and all tokens fall within the price tier above:

Input cost = 10,000 / 1,000,000 × 3.25 = 0.0325 inference credits
Output cost = 2,000 / 1,000,000 × 19.5 = 0.039 inference credits
Total = 0.0715 inference credits

For actual calculations, use the final prices in the model marketplace, which already include the multiplier. For cached requests, check whether the protocol's input count includes cached tokens and which cache read/write prices apply, so overlapping counts are not added twice. Refer to the usage logs for actual charges.

Inference credits are not CNY amounts. Check the amount payable and the allowance received on the top-up/subscription page; transaction fees are described below.

Purchase Methods

Both are bought on the top-up/subscription page.

Subscription

Usually cheaper, valid for 30 days. Eight tiers are currently on sale, from lowest to highest: Lite, Lite+, Plus, Plus+, Pro, Pro+, Max, Max+. All have a monthly allowance with no daily or weekly limit. Check monthly allowances and prices on the top-up/subscription page. For older plans, refer to the allowances and limits shown in My Subscriptions.

Buying the same subscription again does not add a parallel one; it extends the duration, and the later purchase takes effect once the current one expires.

If your allowance runs out before the plan expires, revoke the exhausted plan in My Subscriptions, then buy a new one on the top-up/subscription page. If you have already bought another period of the same plan and it is queued, revoking the exhausted plan activates the next period automatically.

Several different plans can be active at once. The allowance expiring soonest is consumed first, and the account balance is charged only when no plan is available.

Usage-Based

Billed by actual usage, with no time limit.

Payment and Fees

Top-ups are processed through Stripe. Alipay and other former channels have all been migrated.

Each transaction carries a fixed fee of 2.7 CNY. The variable fee (2.2% of the transaction total) has been waived. The fixed fee does not scale with the amount, so larger single top-ups spread that cost further.

Plan Name Mapping

Former and current names

Plan names were standardised. Only the names changed - prices, allowances, entitlements, and active services are unaffected. Existing plans display under the new names automatically.

Former name	Current name
Basic	Lite
Standard	Lite+
Pro	Plus
Max	Plus+
Ultra	Pro
Elite	Pro+
Enterprise	Max
Ultimate	Max+

Related Pages

• Core Concepts - groups, account pools, and billing order
• Model marketplace - prices for every model
• Top-up/Subscription - buy inference credits
• Invoices - how invoices are handled
• Referral Rewards - earn credits by inviting others