Frequently Asked Questions

CheapestToken is a prepaid gateway to discounted LLMs. One key gives you every model we route, all behind a single OpenAI-compatible endpoint — no per-model accounts or keys.

Sign in, top up your wallet, and mint a key on the keys page. Then point your OpenAI SDK's base_url at https://api.cheapest-token.com/v1 and send your first request.

Yes — it is a drop-in replacement. We serve the same /v1/chat/completions and /v1/models routes, so the official OpenAI Python and Node SDKs work with only a base URL and key change.

Usage is deducted from your wallet balance on every call, so there is never a surprise bill at the end of the month. When your balance hits $0, calls stop until you top up.

Prices are USD per 1M tokens. The struck-through price is the upstream list price and the highlighted one is what we charge — the Discount column is your saving.

Cache write is the small charge for storing a prompt prefix, cache read the cheaper charge for reusing it. Repeated prompts and long system messages get cheaper on every hit.

When your input crosses a model's short-context limit, the whole call bills at the long-tier price. Stay under the limit and the short-tier price applies.

The models page is the human catalog. For code, GET /api/v1/models returns context windows and full pricing, and /v1/models serves OpenAI-shaped ids for SDK compatibility.

Console → Usage breaks calls down per model and per day, including cache read and write tokens and cost. Numbers land about a minute after the call.

Create and revoke keys under Console → Keys; revocation is immediate. Treat keys like passwords and rotate them whenever you like.

The security & contact page covers reporting and responsible disclosure, or write to support@cheapest-token.com — a human reads everything.