Short answer: yes, on and off. OpenAI’s own status page logged a new incident at 12:38 AM ET today, September 1, flagging elevated errors in ChatGPT conversations for Free and Go plan users. That’s on top of a bigger outage just yesterday that knocked out ChatGPT Work for a chunk of Plus subscribers. If your chat is hanging, timing out, or throwing a network error right now, you’re not imagining it — and you’re not alone.
We checked OpenAI’s status page directly instead of just trusting Twitter panic, so here’s what’s confirmed, what’s not, and what to actually do about it.
What’s Actually Down Right Now
As of early September 1, OpenAI is investigating elevated errors in ChatGPT conversations for Free and Go plan users — the free tier and the lower-cost Go plan, specifically. That’s a different issue from yesterday’s problem, which hit ChatGPT Work on web and mobile, mostly for Plus subscribers.
Here’s the split that most trackers gloss over:
- Core chat (sending a message, getting a reply) — intermittently affected depending on plan and region
- ChatGPT Work (the agentic task-running mode) — this is the feature that’s been the most fragile lately, with users reporting it “hanging forever” with zero response
- Document uploads and file processing — reported as slow or failing during the worst of yesterday’s incident
- Login and account access — mostly stable this week, unlike some past outages
If you’re on a Plus plan and Work mode is spinning with no output, that’s the known issue. If you’re on Free or Go and getting flat-out errors mid-conversation, that’s the fresh one from today.
This Is the Third Incident in About a Week
That’s the part worth paying attention to. OpenAI’s status history shows:
- August 27 — Elevated error rates in Workspace Agents and ChatGPT Work, fully recovered same day
- August 31 — A bigger outage starting around 11:04 AM ET, hitting Work mode hard for Plus users, with reports rolling in from the US, UK, Mexico, Greece, Japan, India, and Australia
- September 1 — A new, separate incident affecting Free and Go plan conversations, still under investigation as of this writing
Three incidents in six days isn’t random bad luck. It lines up with something status-tracking sites have flagged for a while: independent monitor IsDown has logged roughly 190 OpenAI incidents since October 2025 — an average of 19 a month — and says the typical fix takes about 448 minutes, or just under 7.5 hours. ChatGPT Work, OpenAI’s newer agentic feature, keeps showing up as the common thread in the last few incidents, which suggests the growing pains are concentrated there rather than in basic chat.
We’re not going to pretend we know OpenAI’s internal infrastructure. But the pattern — Work-mode failures, then a plan-specific chat issue days later — reads like a team pushing hard on a feature that isn’t fully stable yet, not a one-off hiccup.
Is It Down for Everyone, or Just You?
Before you assume it’s a you-problem, run through this in order:
- Check OpenAI’s official status page (status.openai.com) first. It’s more reliable than Downdetector because OpenAI posts it directly, not based on crowd-sourced complaints.
- Try a hard refresh or a different browser. Cached sessions sometimes throw errors that a clean reload fixes, even mid-outage.
- Switch networks if you can. Some reports (like the Mexico and Greece complaints from yesterday) suggest regional routing issues on top of the core outage — a different network sometimes routes around it.
- Check the app vs. the web version. During Work-mode incidents, plain chat on the web version has often kept working even while the app or Work mode hangs.
- Wait it out if it’s a confirmed incident. Based on OpenAI’s own average recovery time, budget for a few hours, not a few minutes.
If none of that helps and the status page shows “investigating,” it’s not you. It’s them.
What to Do While You Wait
Losing ChatGPT mid-task is annoying, especially if you’re mid-workflow. A few honest options:
- Claude and Gemini both work as stopgaps for most everyday chat tasks — drafting, summarizing, quick research — if you need something right now.
- Save your prompt before retrying. Repeated retries during an active incident tend to just add to the error pile without getting you a response any faster.
- Don’t re-upload large documents repeatedly. If uploads are the specific failure point, hammering “retry” on a big file just wastes your time.
- Check back on the status page rather than refreshing ChatGPT itself every few minutes — it’s a lighter page and gives you the real update timeline.
Why This Keeps Happening
OpenAI doesn’t publish detailed post-mortems for most of these — the status updates tend to stay generic (“we’ve identified the issue,” “we’re monitoring the recovery”), which is standard practice across most major AI platforms, honestly. Anthropic and Google have both had their own outages this year too.
What’s different here is frequency and which feature keeps breaking. ChatGPT Work is newer and does more — running multi-step tasks instead of a single request-response — which means more moving parts and more ways for something to fail under load. That tracks with the August 27 and August 31 incidents both centering on Work mode specifically, before today’s issue shifted to Free/Go plan conversations instead.
Bottom Line
ChatGPT is having a rough week. If you’re seeing errors right now, check status.openai.com before anything else, figure out whether it’s Work mode or core chat that’s affected, and plan for a wait measured in hours if it’s confirmed. This page will get updated as we learn more.
FAQ SECTION
Q1: Is ChatGPT down right now? A1: Partially. As of early September 1, 2026, OpenAI is investigating elevated errors in ChatGPT conversations for Free and Go plan users. This follows a larger outage on August 31 that hit ChatGPT Work for Plus subscribers. Check status.openai.com for the live status.
Q2: Why is ChatGPT down today? A2: OpenAI hasn’t given a detailed cause, only confirming “elevated errors” under investigation. The pattern across the last three incidents points to instability in ChatGPT Work, the agentic task-running feature, rather than a single isolated bug.
Q3: Is ChatGPT down for everyone, or just me? A3: Check OpenAI’s official status page first — it’s more accurate than social media panic. If it shows an active incident, it’s not just you. If it shows all-clear, try a hard refresh, a different browser, or a different network before assuming it’s your account.
Q4: How long do ChatGPT outages usually last? A4: Based on tracking from independent monitor IsDown, the average OpenAI incident takes about 448 minutes — just under 7.5 hours — to fully resolve, though many smaller issues clear in under an hour.
Q5: What’s the difference between ChatGPT Work being down and ChatGPT itself being down? A5: ChatGPT Work is OpenAI’s agentic mode for running multi-step tasks. It’s failed independently of core chat in recent incidents — meaning you might be able to send normal messages just fine even while Work mode hangs or times out.
Q6: What should I do if ChatGPT is down and I need something done right now? A6: Save your prompt so you don’t lose it, avoid repeatedly retrying (it won’t speed anything up), and use Claude or Gemini as a temporary stopgap for anything time-sensitive.
