Skip to content

Simple, transparent billing

Choose the access pattern that fits.

Token Plans lead the way for predictable model-specific access. PAYG keeps your account flexible across the catalog.

per 1M tokens

Payments & funding

Learn more in the docsReseller program

PAYG vs Token Plan: a worked example

Both funding sources buy the same models through the same API key. The difference is how you pay: PAYG charges each request from a shared balance at the model's per-token rate, while a Token Plan prepays a fixed token allowance for one model. The numbers below are calculated from the live catalog when this page renders.

PAYG balance

Best when you use several models or traffic is irregular. One balance funds every model and you are charged only after a request succeeds.

Claude Opus 4.5 at $0.2500 per 1M input and $1.2500 per 1M output tokens: the workload costs $0.5000 (the same workload at the official list price: $10.0000).

Token Plan

Best when one model carries most of your traffic. A plan adds a model-specific allowance; input and output tokens both draw it down.

The $2.00 Claude Opus 4.5 plan grants 16,000,000 tokens — enough for 13 runs of this workload, about $0.1500 of plan value per run.

Workload used in both columns: 1,000,000 input tokens + 200,000 output tokens on Claude Opus 4.5. Per-leg costs are rounded up to the next micro-dollar, as billing does; your real charge is the one recorded in your usage history.

On every request zcloud spends the model's Token Plan allowance first and falls back to PAYG only for the part a plan cannot cover. With neither, the API returns 402 insufficient_balance and nothing runs.

Token Plans are tied to one model. Your plan balance is used first, then PAYG covers the remainder.

Learn more: Compare models · OpenAI-compatible quickstart · Learn more in the docs

Sign in to buy →