ChatGPT, Gemini, Grok and Claude down as AI services experience outage

Global artificial intelligence services are experiencing major outages according to users.

Those who use ChatGPT and parent company OpenAI, Anthropic's Claude, Elon Musk and X's Grok, and Google Gemini have reported issues on Thursday afternoon (3 September).

The artificial intelligence issue was reported on DownDetector shortly after 3pm BST, with 40,000 users clocking OpenAI services not working as expected.

"It came back up and now it's not working again for me," one user wrote.

Claude has roughly 1,500 problems reported in the same timeframe, with Elon Musk's Grok reporting a similar number.

Google Gemini issues in the hundreds have been noted on DownDetector.

Gemini is down (Thomas Fuller/SOPA Images/LightRocket via Getty Images)

In updates to users, OpenAI said: "We are investigating the issue for the listed services."

This was followed by the company saying it had 'applied the mitigation and are monitoring the recovery'.

According to OpenAI's status page, the following services are impacted: Conversations, Login, ChatGPT Work, Codex in ChatGPT Desktop, Compliance API, Search, File uploads, Voice mode, GPTs, Image Generation, Deep Research, Agent, ChatGPT Atlas, Sites, and Connectors/Apps.

The outage comes shortly after the AI firm confirmed its next major AI model, Astra, is being released soon.

Anthropic, which runs Claude, said via its status page that the following services are impacted: Mythos/Fable 5.1, Mythos/Fable 5, Opus 5, Opus 4.8 and Opus 4.6.

The company said it was 'continuing to work on a fix for this issue' before adding 'we still experiencing degraded performance' in an update shortly before 5pm.

In a follow up update shortly before 4pm, Anthropic said: "The only affected models right now are Opus 4.8 and Opus 5. The rest of the models have recovered to baseline error rate."

Claude is down (Thomas Fuller/SOPA Images/LightRocket via Getty Images)

The Google AI Studio status page reported that its Gemini API had been encoutring problems serving recently created API keys, including through OpenAI-compatible libraries.

The issue hit all services at roughly the same time on Thursday afternoon, with speculation growing as to how and why such a co-ordinated outage could have occurred.