Enterprise
GPUs reserved for you, on terms procurement can sign.
The same API and model IDs you call today, on replicas reserved for you and pinned to a region: H100 and H200 in all 3 regions, B200 in us-east and us-west. A 99.95% uptime SLA, SSO, and SOC 2 Type II, ISO 27001 and GDPR, with a DPA.
- SLA
- 99.95%
- monthly uptime on dedicated
- GPUs
- H100 · H200 · B200
- from $2.49 per GPU-hour
- Regions
- us-east · us-west · eu-west
- Virginia, Oregon and Ireland
- Compliance
- SOC 2 Type II · ISO 27001 · GDPR
- Reports shared under NDA, usually the same day.
Capacity
Dedicated capacity, region by region
Replicas reserved for you: H100 and H200 in all 3 regions, B200 in us-east and us-west. Pin each one to a region where its GPU runs; EU traffic can stay in eu-west on H100 and H200.
| GPU | $ / hrPer GPU-hour | us-east | us-west | eu-west |
|---|---|---|---|---|
| H100 | $2.49/hr | Available | Available | Available |
| H200 | $3.39/hr | Available | Available | Available |
| B200 | $5.49/hr | Available | Available | Not offered |
- Billing
- Hourly rates, billed by the second
- Throughput
- Dedicated endpoints have no per-key request limit: throughput is set by the replicas you reserve.
- Contract
- Committed-volume rates on serverless and dedicated. Invoiced monthly in USD.
- Other models
- Models without a ready replica are set up on request.
Security
Security and compliance
What a security questionnaire asks, what you get, and the evidence behind each answer.
| Area | What you get |
|---|---|
| SOC 2 Type II | Audited every year by an independent firm.Reports shared under NDA, usually the same day. |
| ISO 27001 | Certified information security management, covering every region.Certificate on request. |
| GDPR | EU traffic can stay in (Ireland).A DPA with standard contractual clauses is available on every plan. |
| SSO and SCIM | SSO with SAML 2.0 and SCIM, with audit logs.Works with Okta, Microsoft Entra ID and Google Workspace. |
| Audit logs | Audit logs for every key.Exported as JSON or streamed to your SIEM. |
| Data retention | Zero retention by default.Never used for training. Request logging is opt-in, kept 30 days. |
| Encryption | TLS 1.3 in transit, AES-256 at rest.Covered by the SOC 2 Type II report. |
SLA
99.95% uptime on every dedicated endpoint
On dedicated endpoints, 99.95% a month allows about 21.9 min of downtime. Below that, a share of the month's dedicated bill comes back as credit.
| Monthly uptime | Service credit |
|---|---|
| Below 99.95% | 10% |
| Below 99.5% | 25% |
| Below 99% | 50% |
- Commitment
- 99.95% monthly uptime, with service credits
- Scope
- Every dedicated endpoint is covered, self-serve or under contract.
- Measured
- Per endpoint, each calendar month, from our edge in every region
- Credits
- Applied to the next invoice, with no ticket to file
- Maintenance
- Announced 7 days ahead and run one region at a time
- Serverless
- No SLA. 99.99% API uptime over the last 90 days.
Support
Support that matches what you run
First-response targets by tier. Production down means requests are failing or an endpoint is unreachable.
| Tier | OutageProduction down | OtherEverything else |
|---|---|---|
| StandardEvery accountEmail, at support@weir.example | 4 hbusiness hours | 1 business day |
| DedicatedAny dedicated endpointEmail, plus an on-call page for outages | 1 h24/7 | 8 business hours |
| EnterpriseA contractA named engineer and a shared Slack channel | 15 min24/7 | 4 business hours |
Business hours are 9:00 to 18:00 Pacific, Monday to Friday.
Contact
Talk to an engineer
Four questions, about a minute. Volume and use case are what dedicated capacity is sized from.