AI status
Claude, OpenAI and Kimi have problems right now.
Read from each one's own status page every minute · checked
Downtime, last 90 days
Elevated errors on claude.ai, the Claude API, and Claude Code
We are investigating reports of degraded performance affecting a variety of surfaces including claude.ai, the Claude API, Claude Code, connectors (including Claude Docs and Claude Sheets), Claude Security and Claude Managed Agents.
Elevated errors affecting the Microsoft 365 connector
This issue is caused by a Microsoft 365 service incident, and mail, calendar and search tools in the Microsoft 365 connector are still failing for most requests. Claude for Outlook remains affected. You can follow Microsoft's progress at https://status.cloud.microsoft. We will update this post when service is restored.
claude.ai: 99.56% uptime over the past 90 days, 26 days with problems.
Claude API: 99.78% uptime over the past 90 days, 23 days with problems.
Claude Code: 99.71% uptime over the past 90 days, 24 days with problems.
Claude Console: 99.82% uptime over the past 90 days, 13 days with problems.
Elevated errors affecting the Microsoft 365 connector
We’re investigating an issue with Microsoft connectors where new connections may be broken and tools are not working as expected. We’ll share an update when we know more.
ChatGPT: 99.95% uptime over the past 90 days, 49 days with problems.
Codex: 99.94% uptime over the past 90 days, 9 days with problems.
OpenAI API: 99.99% uptime over the past 90 days, 20 days with problems.
Gemini app: 100.00% uptime over the past 90 days, 1 day with problems.
Gemini API (Vertex AI): 100.00% uptime over the past 90 days, 0 days with problems.
grok.com: 99.80% uptime over the past 90 days, 5 days with problems.
Grok API: 99.78% uptime over the past 90 days, 5 days with problems.
Grok in X: 99.83% uptime over the past 90 days, 1 day with problems.
DeepSeek API: 99.89% uptime over the past 90 days, 16 days with problems.
DeepSeek chat: 99.74% uptime over the past 90 days, 10 days with problems.
Mistral API, Le Chat and Vibe: 99.73% uptime over the past 90 days, 20 days with problems.
Kimi app: 99.99% uptime over the past 90 days, 13 days with problems.
Kimi API: 99.94% uptime over the past 90 days, 6 days with problems.
Kimi Code: 99.55% uptime over the past 90 days, 24 days with problems.
Your AI is down? Keep going.
Morphy runs Claude Code, Codex, Gemini and more on the plans you already have. When one is down, switch the chat to another: click the model button under the text box, then Provider. Your next message takes the conversation along, and the new AI carries on in the same folder. How switching works.
Download Morphy, free for Windows, Mac and Linux.
Where this comes from
Each provider runs its own status page; this one reads them every minute. Uptime is worked out the way those pages do it: a full outage counts fully, a partial one at 30%, and slow not at all. A bar is a day, in the colour of its worst moment; point at one to see what happened.
Gemini, Grok and Mistral only list incidents, not each part's minutes, so there an incident counts by how bad it says it is: critical as a full outage, major as a partial one, minor as slow. Mistral's incidents don't say which part they hit, so it's one row. Gemini comes from Google's Workspace dashboard (the Gemini app) and its Cloud one (the Gemini API on Vertex AI); the Gemini API that AI Studio and Gemini CLI use has no feed to read, see aistudio.google.com/status.
People at each company update those pages, so a problem can be a few minutes old before it shows.