AI provider status — one board, 9 providers

9 of 9 monitored providers operational, 0 with an active incident, plus 2 we can't probe (listed below). Each card links a "how to poll this status API yourself" guide.

Snapshot: Aug 6, 2026 6:50 PM ET · Statuses shown are build-time snapshots — they refresh on each site rebuild. Live auto-refresh ships post-launch.

Unmonitorable from our probe

These providers block or fail our probes — that says nothing about whether their API is up. Check their pages directly.

xAIUnmonitorable

Unmonitorable from our probe — check directly: https://status.x.ai

Probe error: HTTP 403 (Cloudflare challenge; blocks non-browser clients)

DeepSeekUnmonitorable

Unmonitorable from our probe — check directly: https://status.deepseek.com

Probe error: TLS connection reset by peer (host unreachable from this network)

Model retirements

129 announced model deprecation / retirement dates across OpenAI, Anthropic, Google Vertex AI and Azure OpenAI — filterable, cited to each provider's own docs. See the full table →

Frequently asked questions

How often does this status board refresh?

It's a build-time snapshot, not a live poll: every status was fetched from the provider's own status endpoint at Aug 6, 2026 6:50 PM ET and refreshes each time the site rebuilds. The snapshot timestamp is printed at the top of the board. Live auto-refresh ships post-launch.

What does “Operational” mean on this board?

It means the provider's own status feed reported no active incident (indicator “none”) at snapshot time. It is the provider's self-reported status — not an independent probe of API latency or error rates.

Why can't you monitor xAI or DeepSeek?

xAI's status page returns HTTP 403 (a Cloudflare challenge) to non-browser clients, and DeepSeek's status host resets TLS connections from our network. Both are listed as unmonitorable rather than guessed — an absent probe is not an “operational” badge.

Why doesn't polling /api/v2/status.json work for every provider?

Only 5 of the 9 monitored providers use Atlassian Statuspage's standard endpoint. The rest use Instatus, BetterStack, Google's incidents feed, or no JSON endpoint at all — each provider guide documents the endpoint and format that actually works.