The percentage of time a system is operational and accessible, often defined in SLAs. Availability commitments for AI-dependent systems often account for model-specific issues (capacity, rate limits, provider outages).
See: Business continuity; Service Level Agreement / Service Level Objective; Uptime