Wired reports Claude and Grok outages hit within four minutes

• Claude and Grok failed within four minutes of each other, then ChatGPT and Codex followed 73 minutes later, according to a Wired excerpt.
• OpenAI attributed its ChatGPT and Codex disruption to a routing error, while SpaceXAI attributed Grok's to an outage at its Memphis compute center; the same Wired excerpt says Anthropic had not disclosed Claude's cause.
• Claude's incident spanned several named models, while coding clients saw endpoint failures: the Claude status screenshot lists the affected models and a Codex error screenshot captures an unexpected status 404 response.
• Recovery was staggered, with unclebobmartin's update reporting Codex back while Grok still returned an elevated-latency message.

A ChatGPT 404 screenshot captured the web app returning a page-not-found error, while an OpenAI gateway-timeout screenshot showed a Cloudflare 504. The picture was uneven enough that synthwavedd reported other services working while ChatGPT misbehaved, even as olvrgln asked whether everything was down.

Early accounts did not describe one uniform failure. omarsar0 reported both Codex and Claude unavailable, while unclebobmartin described Codex and Grok as degraded.

ChatGPT then joined the visible outage: cedric_chee posted that it was down, and unclebobmartin was asking Grok for an ETA. Both OpenAI and Anthropic acknowledged elevated errors, as recorded by a Bloomberg screenshot.

The near-overlap drove immediate speculation about shared infrastructure. The publicly stated causes that followed point instead to provider-specific incidents.

Anthropic's status page named five affected model entries:

• Mythos/Fable 5.1
• Mythos/Fable 5
• Opus 5
• Opus 4.8
• Opus 4.6

OpenAI failures crossed several request paths. A Codex client request failed at /backend-api/codex/responses in the Codex error screenshot; the ChatGPT 404 screenshot shows the main web endpoint failing; and the gateway-timeout screenshot shows a separate 504 at Cloudflare's edge.

The postmortems available so far split three ways. OpenAI's account to The Register says a routing error made ChatGPT and Codex unavailable for some users. SpaceXAI said Grok's problem was an outage at its Memphis compute center.

Before that attribution emerged, LLMpsycho had pointed to an xAI datacenter failure. Anthropic reported it had identified and fixed its issue, but an Ars Technica reconstruction says the company did not publicly describe the cause.

The Wired report found that neither OpenAI nor Anthropic cited a shared external provider. The overlap remains a coincidence in the public record, not a confirmed common-cause event.

Codex recovered while Grok was still degraded, as unclebobmartin's update recorded. Later, unclebobmartin said he had switched back to Codex because Grok was down.

OpenAI marked the incident resolved after degradation across 15 ChatGPT components and four Codex components. Its status notice added one concrete aftereffect: some Codex remote-control users may need to pair their mobile device again.