What availability means on AI Stats
Provider availability means AI Stats has a model-provider mapping for a specific API route, capability surface, or gateway-executable pairing.
It does not guarantee perpetual uptime, universal account access, or identical commercial terms across every region and customer tier.
How mappings are stored
AI Stats stores canonical model identifiers separately from provider-specific model identifiers so one model can be tracked across multiple providers without losing provider-specific details.
Provider model rows can also carry capability metadata, prompt-training policy overrides, active-gateway state, modality information, and effective date windows.
Gateway-active versus catalog-visible
A provider-model mapping can exist in the catalog before it is fully active on the public gateway. The catalog is meant to show coverage and operational context, while gateway state signals whether that route is available for execution.
That distinction matters because some models appear in provider documentation before they are broadly enabled, stable, or routable through a common interface.
When provider pages change
Provider pages change when model mappings, capability rules, pricing rules, or provider metadata are updated. AI Stats revalidates those pages and can notify external discovery systems such as IndexNow when public URLs materially change.
Because provider coverage is operational data, totals can move as routes are added, disabled, renamed, or recategorized.
Caveats
Availability on AI Stats should be read as a current catalog signal, not a contractual promise. Enterprise allowlists, region locks, quotas, or staged rollouts can still affect whether a route works for a given account.
If a provider page has too little public data to stand on its own, AI Stats may prefer reduced indexing until the page contains enough verified coverage information.