AI API outages on August 22, 2026

UTC day · measured every 5 minutes from 4 regions · method

On August 22, 2026 (UTC) our probes attributed 3 outage(s) to 1 provider(s): Gemini API (Google): 3.

This page is history. Is it down right now?

Live status, updated every 30 minutes: Gemini API (Google) · all AI APIs, last 72 hours

Was the Gemini API (Google) down on August 22, 2026?

Yes, measured. 3 outage(s) attributed to Google: requests to gemini-flash-lite-latest failed from up to 2 of 4 regions at once while no other provider was failing, first at 17:10 UTC, at least 3 minutes observed in total.

UTCWhat failedRegionsDurationFailed probesVerdictError
14:40–14:40requests to gemini-flash-lite-latest2 of 4: South America (São Paulo), US (Central)<5 min2shared with 2 other provider(s) — not attributedThe read operation timed out
17:10–17:13requests to gemini-flash-lite-latest2 of 4: Europe (Germany), South America (São Paulo)≥3 min2provider outageThe read operation timed out
21:20–21:20requests to gemini-flash-lite-latest2 of 4: South America (São Paulo), US (Central)<5 min2provider outageThe read operation timed out
23:00–23:00requests to gemini-flash-lite-latest2 of 4: South America (São Paulo), US (Central)<5 min2provider outageThe read operation timed out

Is the Gemini API (Google) down right now? Live status →

80 single-region failure(s) were also recorded that day (yi-01ai 18, cohere 14, baseten 12, nscale 11, google 6, iflytek 5, baichuan 3, sambanova 2, groq 1, xai 1). A failure seen from one region cannot be separated from a problem on that network path, so none of them is called an outage.

How is an outage told apart from a network problem?

Each provider is probed every five minutes from 4 independent regions. A failure is attributed to the provider only when it is seen from two or more regions at once and no other provider fails in the same minutes; a failure shared with other providers has a cause above all of them and is not blamed on anyone. Durations are lower bounds: we see the first and last failed probe, not the real start and end. Status-page entries are the provider’s own words, shown next to the measurement, not instead of it.

· Last 72 hours ·

Machine-readable: incidents.json · Licence: CC BY 4.0