July 2026 AI Tools Outage & Uptime Report
We continuously monitor 18 major AI tools including ChatGPT, Claude and Gemini. A recap of incidents recorded on providers’ official status pages in July 2026.
Longest outage: Anthropic（28h 26m）
By provider
| Provider | Incidents | Downtime | Worst state | Uptime |
|---|---|---|---|---|
| Anthropic | 16 | 58h 14m | Major outage | 92.17% |
| GitHub | 14 | 29h 44m | Major outage | 96% |
| OpenAI | 8 | 37h 4m | Major outage | 95.02% |
| Cursor | 6 | 12h 6m | Major outage | 98.37% |
| ElevenLabs | 3 | 11h 42m | Partial outage | 98.43% |
| Runway | 1 | 1h 0m | Major outage | 99.87% |
| Cohere | 1 | 51m | Partial outage | 99.89% |
Incident counts and downtime cover only "major-or-worse and resolved" incidents recorded on official status pages (minor latency and unresolved ongoing items are excluded). Uptime is our own estimate based on that, not official vendor figures. Aggregated by calendar month in Japan time (JST).
Highlights of the month
July 2026 saw 48 major incidents totalling 8,901 minutes across 7 providers. Anthropic logged the most (16), followed by GitHub (14) and OpenAI (8).
The longest single entry was Anthropic’s suspension of access to Claude Mythos 5 and Claude Fable 5, which began on June 13 and was resolved on July 1 — 1,706 minutes counted toward July. Note that this was closer to a deliberate suspension than a fault; last month’s report excluded it as an unresolved item.
The most impactful incident that actually started in July was OpenAI’s "image generation unavailable in ChatGPT", which lasted roughly 16 hours from July 21 19:36 to July 22 11:50 (JST). The same day also saw elevated error rates for image generation on the API side (about 2h46m), making image generation the clear weak spot of the month.
OpenAI logged a second long incident: for about 14 hours from the early hours of July 18, Enterprise users without Codex permissions could not use the new ChatGPT app.
On the developer side, GitHub Actions stood out. Delays starting runs lasted about 9 hours on July 9, and another Actions incident ran about 5 hours on July 20. GitHub was the second most frequent provider with 14 incidents.
Anthropic’s 16 incidents were mostly announced per model, spanning Opus 5, Sonnet 5, Haiku 4.5, Fable 5, Mythos 5 and Opus 4.8/4.1. Each typically resolved within one to five hours.
In voice AI, ElevenLabs recorded WhatsApp outbound message failures (about 5 hours) and Speech Engine conversation failures (about 4h20m).
Trends & takeaways
This month showed clearly how an upstream outage cascades downstream. On July 29, about nine minutes after Anthropic announced elevated errors across all models, Cursor announced "elevated errors affecting Anthropic models". Several of Cursor’s six incidents originated with its model providers — a reminder that dependence on upstream APIs, rather than the IDE or cloud features themselves, is the availability risk.
Some days clustered across providers. On July 16, Cursor, GitHub and Anthropic all logged major incidents; on July 21 it was OpenAI, Anthropic and GitHub. Looking at a single vendor you might wonder whether the problem is on your end — seen side by side, it is clearly simultaneous.
Minor degradations (latency-level) numbered 101, more than double the 48 major incidents. Most of the "it just feels slow" experience falls into this category.
Anthropic shows the lowest uptime at 92.17%, but partly because it publishes incidents at a fine-grained, per-model level. Providers that disclose in more detail look worse in the numbers — worth discounting when comparing.
How to read this report
Incident counts and downtime include only incidents on each provider’s official status page that were both "major or worse (effectively unusable)" and already resolved (calendar month, Japan time). Minor degradations such as latency, and unresolved ongoing items, are excluded from the totals so the figures aren’t inflated beyond real impact.
This month’s total includes 1,706 minutes from Anthropic’s carried-over model suspension, resolved on July 1. As that was not a genuine fault, subtracting it gives a more realistic picture of the month.
ElevenLabs’ SIP call failures, which began on July 31, were still unresolved at month end and are therefore not counted in downtime.
Uptime is our own estimate based on that major-outage time — not official SLA figures.
Each service’s "outage history" page lists individual incidents with title, severity, duration and the official link.