Users report issues with Claude: Is it down?

Claude users experienced a fresh wave of trouble Wednesday as the artificial intelligence platform briefly went down, triggering more than 2,000 outage reports and elevated errors across multiple Anthropic models.

The outage affected users across the app and web console, with Anthropic warning of “elevated errors across multiple models” and signaling that further service updates may follow as they continue investigating.

In a statement to Newsweek, an Anthropic spokesperson confirmed that service had been fully restored.

"Service has been fully restored across Claude.ai, Claude Code, and the Claude API. We know how much people rely on Claude, and we're grateful for everyone's patience while our team worked to resolve the issue," they said.

Earlier today, the company had appealed for "everyone's patience as we restore full service."

Downdetector logs show issues began shortly after 3:30 p.m. EDT and were resolved within the hour, with roughly half of all complaints tied to Claude Code.

Claude wasn’t working because the platform experienced elevated error rates across multiple models, triggering a major outage affecting key services such as Claude Code, the Claude API, and parts of Claude Chat.

The problems began just after 3:30 p.m. EDT, with more than 2,000 outage reports on Downdetector. Roughly half of the issues involved Claude Code, though users also struggled to access the app and console.

Anthropic’s status page showed multiple services in “major outage” despite their typical 99%‑plus uptime over the past 90 days.

What Is the 'Capacity Constraints' Error on Claude?

The “capacity constraints” error on Claude appears when the system is under heavy, platform‑wide demand and Anthropic’s servers temporarily can’t handle additional requests. It’s essentially Claude’s way of signaling that overall traffic—not your personal usage—is overwhelming available compute.

Claude returns this message when server load triggers an HTTP 529 status code, indicating platform overload, unlike an HTTP 429 rate limit, which means you exceeded your individual quota.

During peak global usage, Anthropic may prioritize paid tiers or high‑value workloads, briefly throttling free or lighter operations.

Most users see the error clear within minutes. Anthropic recommends waiting rather than repeatedly resending the same prompt, since rapid retries can worsen queue congestion.

How Many Outages Has Claude Had This Month?

Claude has faced a steady drumbeat of service disruptions throughout July, with repeated spikes in “elevated errors” across its major models and tools. The pattern shows the platform struggling under load or encountering system‑wide faults, often affecting multiple generations of Opus, Sonnet, Haiku, and newer experimental models like Fable and Mythos.

Across the incidents, Anthropic repeatedly reported “elevated errors across all models,” including on July 29, when the company logged an unresolved incident: Elevated errors across all models. The day before, the platform saw failures on Opus 5 and Haiku 4.5, noting that errors have returned to baseline after multiple investigation updates.

Earlier in the week, outages cascaded across Fable 5, Sonnet 5, Haiku 4.5, Mythos 5, and Opus 5, with Anthropic repeatedly stating it was “investigating elevated errors” and “implementing fixes” as issues rippled through the system.

Some disruptions extended beyond model performance. On July 24, users of Claude’s Microsoft Office add‑ins were unable to access features until Anthropic identified the cause and issued a fix, noting: "Users who installed the following add-ins may experience issues accessing the add-in functionality."

Contact Newsweek editors on this story: Samantha Beech and Sam Wilson.