Service Level Agreement

Last updated September 17, 2026

This SLA describes the availability of the Lingyuns gateway and GPU instances, and the remedies available when we miss the target.

1. Gateway availability

Target: 99.9% monthly availability for the inference gateway, measured per calendar month across all successfully authenticated requests except those failing due to your content, rate limits or upstream model deprecations.

2. Service credits

Below 99.9%: 10% of the affected month's token spend credited. Below 99.0%: 25%. Below 95%: 50%. Credits are applied to your balance within 30 days of a verified claim.

3. GPU instances

GPU instances carry a 99.5% monthly availability target counted from the moment the instance reports ready. Unplanned downtime is credited pro-rata in minutes.

4. Exclusions

Scheduled maintenance (announced 48h ahead), force majeure, your software stack, network paths outside our control, and free trial or beta capacity are excluded from credit calculations.

5. Claims

Open a claim within 30 days of the incident by emailing [email protected] with request IDs and timestamps. Approved credits are the sole remedy under this SLA.