Pricing
Pay for tokens. Nothing else.
There is no subscription, no seat count and no minimum commitment. Add credit to your balance and it is drawn down request by request at the published per-million rates.
One plan, every capability
Features are not gated behind tiers. The only variable is how many tokens you send.
- Access to every model in the catalog
- Automatic failover across providers
- Streaming and tool calling
- Per-request logs with cost and latency
- Unlimited API keys with per-key limits
- Daily usage rollups and monthly invoices
How billing works
Each request is priced as input tokens times the input rate plus output tokens times the output rate. Reasoning tokens are billed at the output rate, exactly as the upstream provider bills us, and are reported separately so you can see what thinking cost. Requests that fail before producing tokens are never charged.
Answers before you sign up
For chat completions, yes. The request and response bodies, the streaming chunk format and the error envelope all match OpenAI, so official SDKs and anything built on them work without modification.