Cloud Horizon AI / Pricing
Pricing that maps to how procurement actually buys.
A flat plan fee for the platform, metered tokens per model on top. No bundled credits, no opaque conversion ratios, no per-seat math games. The PDF you send to finance is one page.
Personal
For solo builders who want one EU gateway and predictable usage pricing.
500 requests / month included
Then metered token pricing per model.
- · All open-weights models in the catalog
- · EU region pinning (eu-ams-1, eu-fra-1)
- · 30-day default audit log retention
- · PII redaction toggle per request
- · Email support, 2 business day SLA
Team
For teams that need workspace billing, audit trails, and signed DPA.
5,000 requests / month included
Metered token pricing, 8% volume discount above 100k requests / month.
- · Everything in Personal
- · Workspace seats and SSO (SAML, OIDC)
- · Per-tenant audit tag isolation
- · Standard EU DPA, signed within one business day
- · Audit logs streamable to your S3 bucket
- · Slack support channel, 4-hour weekday response
Enterprise
For regulated workloads and procurement teams that need contracts.
Custom volume commit
Net-30 invoicing, EUR or USD, multi-year discounts available.
- · Everything in Team
- · Customer-managed KMS keys (BYOK)
- · Dedicated capacity in eu-ams-1 or eu-fra-1
- · Negotiated DPA, security questionnaire, SIG/CAIQ pre-filled
- · Named technical account manager
- · Quarterly business review and roadmap input
Token economics by model
Per-model rates in USD per million tokens. Same rates for every plan tier; the plan fee only buys the platform around the model.
| Model | Input ($/Mtok) | Output ($/Mtok) | Notes |
|---|---|---|---|
| Kimi K2.5 | $0.55 | $2.20 | Long-context champion, 1M tokens context. |
| GLM 4.6 | $0.50 | $1.50 | Best price-to-quality for general workloads. |
| Qwen 3 Coder | $0.30 | $1.20 | Code generation and review specialist. |
| MiniMax M2.5 | $0.40 | $1.60 | Multilingual and reasoning balance. |
| Llama 3.3 70B | $0.20 | $0.80 | Cheapest option, fully open weights. |
| Mistral Large 3 | $0.70 | $2.10 | Strongest function calling. |
- ·Prices in USD per 1 million tokens. EUR conversion is 0.92x at the time of invoice.
- ·Cached input tokens, when supported by the model, are billed at 25% of the input rate.
- ·Zero-retention requests have a 5% surcharge to cover the in-memory inference path.
- ·No surcharge for region pinning, audit tags, or PII redaction. Those are part of the platform.
Estimate a month of usage
Three knobs: model, requests per month, average input and output tokens per request. Returns the all-in monthly cost for the Team plan. Conservative estimate, no assumptions about caching or zero-retention surcharge.
Estimated monthly cost (Team plan)
$0
- Plan fee
- $0
- Input tokens
- $0
- Output tokens
- $0
- Included requests
- 5,000 / month
Estimates are advisory. Actual invoices reflect tokenizer counts at the gateway, not application-level estimates.
Pricing FAQ
The questions procurement asks before they sign.
Is the per-request bundle metered separately from tokens?
Yes. Every paid plan includes a base request allowance plus token usage at the per-model rate. We do not bundle them into a single opaque "credit" unit because procurement teams need to model both dimensions independently.
Can I switch currency mid-contract?
On Personal and Team, billing currency follows the workspace setting and can be flipped any time. On Enterprise, the currency is locked into the order form and FX is settled at signing.
Do unused included requests roll over?
No. We considered rollover but it makes capacity planning harder for both sides. If your usage is bursty, the Team plan with overage pricing is usually the better fit than stockpiling Personal credits.
How do you handle VAT?
EU customers pay 21% Dutch VAT unless a valid VIES VAT number is supplied. UK customers pay 20% UK VAT. Outside the EU and UK, no VAT is charged. Stripe handles the reverse-charge mechanics.
What happens if I hit the included request cap mid-month?
You get a soft warning at 80% via email and Slack webhook. At 100%, requests continue at the metered rate; we never hard-stop a Team or Enterprise account without consent. Personal plan can opt into a hard stop.
Are there proof-of-concept credits?
Yes. Team plan trials get 25,000 free requests for 30 days, no credit card required. Enterprise pilots are scoped per opportunity, typically 90 days with a usage cap negotiated up front.
Ready when you are
Procurement-friendly EU AI infrastructure
Join the waitlist for Personal and Team. Email [email protected] for Enterprise pilots.