Pricing
Pay for what you use. Per-token model pricing and per-unit add-ons, with no minimum.
Model pricing
What each model costs
Priced per token, billed to the minor unit. No subscription, no minimum spend.
The price list didn't load. Try again in a moment.
The model catalogue is empty right now.
We haven't set public rates for these models. Get in touch and we'll quote you directly.
| Model | Input | Output |
|---|
Prices are per 1,000,000 tokens.
Add-ons
Tools and add-ons
Metered separately from model tokens, only while you use them.
The add-on price list didn't load. Try again in a moment.
Store your own documents and query them for retrieval-augmented answers.
Per GB stored / month, plus per query
Packaged capabilities a model can call while it answers.
Per enabled skill / month
Connect external MCP tool servers to your models.
Per connected server / month
Let a model search the web for current information.
Per 1,000 searches
Run model-generated Python in an isolated sandbox.
Per session-minute
Render HTML to a downloadable PDF document.
Per document
Find your plan
Every plan uses the same API and the same models — the difference is scale.
Get an API key and try the platform before you add billing.
- An API key the moment you sign up
- Rate-limited access for evaluation and testing
- Move to pay-as-you-go anytime, same key
Add credit and pay only for what you use, metered as you go.
- Every model, tool and add-on, priced per unit
- Billed from your balance in real time, no invoice to reconcile
- Requests pause automatically if your balance runs out
Custom throughput and volume pricing for high-usage teams.
- Custom rate limits and dedicated throughput
- Volume pricing across models and add-ons
- Priority support for production workloads