One key, one invoice. You pay for the tokens you use - prepaid, per million. We meet OpenRouter pricing on aggregated models and undercut it on our sovereign models hosted in Canada.
Top up $10, $25, $100 or $500 and draw down as you call. No monthly minimum, no seat fees, no commitment.
Priced per 1M input and output tokens, metered per request. Image models bill per image. You only pay for what you generate.
Sovereign models and aggregated external providers on a single balance and a single statement. No juggling five vendor bills.
Our flagship open models run on our own Canadian hardware, so we can price them below the closed-model incumbents.
Open-weight models we serve on our own NVIDIA hardware in Canada. Your data stays in-country. Because we own the metal, we undercut closed-model pricing - and you get 1M-token context and native tool-calling on the flagship.
GPT, Claude, Gemini and other frontier models, proxied through one modelhaus key. Priced at provider / OpenRouter parity with transparent, at-cost aggregation - so you can mix sovereign and external behind a single endpoint.
Live from our catalog - per 1M tokens (or per image). Updated continuously.
| Model | Type | $ / 1M in | $ / 1M out | Context |
|---|---|---|---|---|
| Loading catalog… | ||||
Prices in USD. Image models bill per image. Aggregated models are billed at provider parity. Sovereign models are hosted in Canada. See the full catalog for capabilities and status.
No. modelhaus is fully prepaid and usage-based - top up a balance and spend it down. No monthly fee, no seat licences, no commitment.
Every request is metered by input and output tokens (or per image for image models) at the model's listed rate and deducted from your balance. Your dashboard shows per-request accounting in real time.
Aggregated external models are billed at provider / OpenRouter parity - you pay what you'd pay going direct, with the convenience of one key and one invoice. The savings come from our sovereign models, which we price below the closed incumbents.
Generous default limits scale with your balance and usage history. Need dedicated throughput or a private deployment? Talk to us.
Calls to sovereign models stay entirely on Canadian infrastructure. Calls you route to aggregated providers go to that provider under our unified key. You choose per request.
Create an account, top up, and call every model through one OpenAI-compatible endpoint. No sales call required.
Get an API key →