Pay for tokens. Not for the plumbing.
Every model is billed at the provider's own per-token rate. Routing, failover and caching are part of the platform, not a line item on your invoice.
Pay as you go
$0plus token usage
Top up credits and spend them on any model. No seats, no commitment.
- All 350+ models, one API key
- Provider token rates, 0% routing markup
- Automatic failover across providers
- 20 requests / second
- Usage dashboard and CSV export
Team
Most popularTBDTo be disclosed
Shared keys, per-project budgets and the routing controls a team needs.
- Everything in Pay as you go
- Per-key and per-project spend caps
- Custom routing rules and model fallbacks
- Prompt caching analytics
- 200 requests / second
- SSO via Google and GitHub
Enterprise
Customannual agreement
Committed spend, private routing and the paperwork your procurement team wants.
- Everything in Team
- Volume discounts on committed spend
- Zero-data-retention routing only
- Region pinning (US / EU)
- SAML SSO, SCIM and audit logs
- 99.9% uptime SLA with support escalation
Estimate the bill before you ship
Set your traffic shape and see what each model would actually cost a month. Switching the model string is the only change your code needs.
$3.00 in · $15.00 out per 1M tokens
Estimated monthly spend
$1,500
$0.01 per request
- Input
- $600.00
- Output
- $900.00
- Routing fee
- $0
Cached tokens are billed at roughly a tenth of the input rate, so a high hit rate on a long system prompt is the cheapest optimisation available to you.
Live price list
Sorted cheapest first, straight from the registry. The full list lives in the model catalog.
Showing the 20 cheapest of 319 language models. Prices are per million tokens, read live from the registry.