ChatGPT Outage Report — April 2026

April 2026 ChatGPT outage report: · 12 incidents · 99.52% uptime · 280ms avg response

Report period: April 2026 (Apr 1 – Apr 30, 2026 UTC).

Executive Summary

During April 2026, ChatGPT had
12 recorded incidents
with an independent-probe uptime of 99.52%
and an average response time of 280ms
.

Key Metrics

Independent probe uptime
99.52%
Average response time
280ms
Recorded incidents
12
Longest outage
—
Checks executed
4623
User reports submitted
3

Most Affected Services

Complete Incident List

  1. GPT-5.1 Codex Max Elevated Error Rate9:56 AMAll impacted services have now fully recovered.

    · codex · Minor
  2. ChatGPT issues while retrieving workspace members9:00 PMAll impacted services have now fully recovered.

    · chatgpt · Minor
  3. Elevated error rates for Responses API11:53 PMAll impacted services have now fully recovered.

    · api · Minor
  4. Some Codex cloud tasks are failing2:30 AMAll impacted services have now fully recovered.

    · codex · Critical
  5. High error rate for Dall-e2:45 AMAll impacted services have now fully recovered.

    · dall-e · Minor
  6. Increased error rate for ChatGPT and API for prompts with images9:12 PMChatGPT and API functionality should now be fully restored.

    · api, chatgpt, dall-e · Minor
  7. Connectors/Apps unselectable in ChatGPT Input Menu7:45 PMConnectors/Apps are again selectable in the ChatGPT Input Menu.

    · chatgpt · Minor
  8. Elevated error rates for ChatGPT users1:03 AMAll impacted services have now fully recovered.

    · chatgpt · Minor
  9. Codex Github Issues2:55 PMAll impacted services have now fully recovered.

    · codex · Minor
  10. ChatGPT availability degraded for free and logged out users 5:27 PMAll impacted services have now fully recovered.

    · chatgpt · Minor
  11. Elevated Codex Error Rate11:03 AMAll impacted services have now fully recovered.

    · codex · Minor
  12. Brief Issue with Image Generation Results2:14 AMAll impacted services have now fully recovered.

    · dall-e · Minor

Methodology

Status is computed by combining the official OpenAI status feed, independent HTTP probes, scraped incident history, and (when available) user reports and regional probes. For the full decision logic and confidence scoring, see our methodology page.

Related