DCAIPredictable gateway costs. Your choice of models.#
Compare plans for your own applications or a portfolio of customer deployments. Connect your own provider account, keep model usage on that account, and pay DCAI for gateway access and routing.
USD monthly prices below come from Gateway documentation dated October 15, 2026. Confirm the final price and availability in Marketplace or your private offer before purchasing. Provider inference, taxes and currency conversion are additional.
Team
$199 / month
Build and operate your team’s AI integrations.
- 5 active keys included · 25 keys maximum
- 200M open-weight routed tokens
- 25M frontier routed tokens
Partner Edition
$499 / month
Manage more keys and larger routing allowances.
- 25 active keys included · 100 keys maximum
- 1,000M open-weight routed tokens
- 100M frontier routed tokens
Estimate your monthly Gateway charge#
Enter tokens in millions, adding prompt and completion tokens. Enter active keys for one subscription billing period. For example, 100 means 100 million tokens.
The estimate applies included allowances across the subscription. It does not predict your model provider's bill, taxes, credits, negotiated discounts, proration or currency conversion. It does not verify capacity: per-key limits and provider quotas still apply. Inputs stay in your browser; no account or API key is needed.
What is included?#
| Capability | Team | Partner Edition |
|---|---|---|
| Keys per subscription, including Default | 25 | 100 |
| Active keys included each billing period | 5 | 25 |
| Included open-weight routed tokens | 200M | 1,000M |
| Included frontier routed tokens | 25M | 100M |
| Tokens per minute, per key | 60,000 | 120,000 |
| Calendar-month token cap, per key | 500M | 2,500M |
| Additional active key | $15 per billing period | $15 per billing period |
| Open-weight overage | $0.25 per 1M tokens | $0.25 per 1M tokens |
| Frontier overage | $0.60 per 1M tokens | $0.60 per 1M tokens |
Both plans provide tenant keys, per-key token budgets and usage CSV exports. Provider and endpoint support varies; review the API reference and provider guides before choosing models. No model inference credits are included in these routing allowances.
How the bill works#
One active key is one metered tenant. A key becomes active when it routes at least one token during the subscription billing period. The Default key counts. One business customer using several keys can therefore produce several active-tenant charges. Unused keys do not incur an active-key overage, but still count toward the key limit unless revoked.
Your provider bills model usage separately. DCAI routing fees are additional to Microsoft Foundry or Google Vertex AI inference charges. DCAI does not resell the underlying model access under these plans.
Model tier affects the routing rate. Map and test deployments before estimating. Untested or unconfirmed deployments can count as frontier. See billing tiers and metering rules.
Budgets are in tokens. Per-key budgets stop requests at the configured token limit; they are not dollar-spend guarantees. Calendar-month caps and subscription billing-period allowances can have different reset dates.
Build directly or deliver a managed service#
For a direct integration, start with the quickstart, then use the production checklist.
For multiple customers, use the MSP & partner guide to plan tenant boundaries, customer reporting and a private offer. Larger capacity, dedicated infrastructure and custom support need a scoped discussion and written terms; they are not included automatically.
Ask about your deployment · Read the detailed billing documentation