> ## Documentation Index
> Fetch the complete documentation index at: https://docs.corvex.cloud/llms.txt
> Use this file to discover all available pages before exploring further.

# Pricing

> Per-token pricing in USD for every model on Corvex. Pay only for what you use — prepaid credits, no minimum spend, no subscription.

# Pricing

Token-based pricing in USD. You pay only for the tokens you send (input) and the tokens the model returns (output). No subscription, no minimum spend, no per-seat fees — the same rates apply to every account.

## Per-million-token rates

| Model          | Input       | Output      | Card                                                                   |
| -------------- | ----------- | ----------- | ---------------------------------------------------------------------- |
| GLM 5.2 FP8    | \$0.75 / 1M | \$2.40 / 1M | [/models/zai-org/GLM-5.2-FP8](/models/zai-org/GLM-5.2-FP8)             |
| Kimi K2.7 Code | \$0.55 / 1M | \$2.20 / 1M | [/models/moonshotai/Kimi-K2.7-Code](/models/moonshotai/Kimi-K2.7-Code) |

Rates are per million tokens. A typical 1k-token chat turn costs a fraction of a cent. See each model card for context window, capabilities, and example prompts.

## How billing works

Corvex uses prepaid credits. Top up once and the platform meters every request against your balance.

* **Buy credits** in any amount once you sign in — credits land in your account immediately and never expire.
* **Auto-reload** keeps your balance above a threshold you set, so production traffic doesn't stall.
* **Usage pauses at \$0** — requests return an `insufficient_credits` error instead of running up a surprise bill.
* **One invoice per top-up**, downloadable from the dashboard for accounting.

## Monthly spend caps

A workspace owner can set an optional **monthly spend cap** — a hard dollar ceiling (for example, `$500/month`) on top of your prepaid balance.

* **Set it once.** The cap lives on the workspace. While usage stays under the cap, requests run normally.
* **Hard stop at the cap.** When the month's spend reaches the cap, inference requests return HTTP `402` with an [`insufficient_credits`](/reference/errors#insufficient_credits-402) error until the cap resets or is raised. There is no soft/overage mode.
* **Resets monthly.** The cap resets at the start of each calendar month (UTC).
* **Independent from token quotas.** A spend cap is dollar-denominated and is separate from per-key token rate limits (which return `429`). New keys are provisioned with **no default token cap** — usage is bounded by your prepaid balance and any spend cap. You can still set a per-key token limit explicitly; whichever limit is hit first applies.
* **Raise it anytime.** Increasing the cap unblocks requests immediately — no waiting for the reset.

## What's included

Every account on the public catalog gets the same platform features at the rates above:

* Every model in the table on Corvex's shared inference fleet.
* OpenAI-compatible `/v1/chat/completions` and `/v1/models` endpoints. A `/v1/embeddings` endpoint exists, but no embedding model is currently deployed.
* Per-key usage metering, rate limits, and structured error responses.
* The Corvex playground and dashboard at [tokenfactory.corvex.cloud](https://tokenfactory.corvex.cloud/app).

## FAQ

### Do I need a contract to get started?

No. Sign in, top up, and call the API. The dashboard handles keys, usage, and invoices end-to-end.

### Is there a free tier or free credits?

New accounts can claim starter credits after sign-up. See the dashboard for the current promotion.

### What about steady, high-volume workloads?

[Corvex Reserved Endpoints](https://corvex.ai/contact) offer dedicated single-tenant GPU capacity at a fixed per-hour rate. Better unit economics once your traffic is predictable.

### When does pricing change?

Rates on this page mirror the platform catalog at [/v1/models](/api-reference/openapi). When a rate changes, the next API response and this page update together, so prices always match what `GET /v1/models` returns.

<Note>
  The rates on this page always match what `GET /v1/models` returns. Call that endpoint any time for the authoritative, up-to-date per-model pricing.
</Note>

Ready to build? [Sign up at tokenfactory.corvex.cloud](https://tokenfactory.corvex.cloud/app/sign-up) to claim your starter credits.

Need a model that isn't listed? [Tell us](mailto:hello@corvex.ai).