Step-by-Step Fix
1. Confirm the exact symptom
Write down the exact message you see when image generation fails — "I cannot generate images," a specific error code, or just no response. Note when it started and whether it affects all prompts or only specific ones.
2. Verify plan and model access
Image generation in ChatGPT requires:
- A Plus, Team, or Pro subscription (not available on free plan)
- The GPT-4o model selected (not GPT-4o mini)
Check the model selector at the top of the chat. If it shows GPT-4o mini, switch to GPT-4o. If you are on the free plan, image generation requires upgrading to Plus.
3. Run the two isolation tests
Test 1 — Incognito window: Open a private browser window, sign in, and attempt image generation.
- Works in incognito → browser extension or local cache issue in your regular browser
Test 2 — Phone hotspot: Switch to mobile data and retry.
- Works on hotspot → your network is filtering OpenAI's image generation endpoints
4. Reset browser and session state
- Sign out and sign back in
- Hard refresh with Ctrl+Shift+R (or Cmd+Shift+R on Mac)
- Clear site data for
chat.openai.comandopenai.com - Disable extensions temporarily — particularly any that modify content or block requests
5. Disable VPN and proxy
VPN endpoints can be blocked for OpenAI's image generation service specifically, even if text generation works fine. Disable any VPN or proxy and retry on a direct connection.
6. Verify account and workspace context
- Confirm you are using the account with the Plus subscription
- If you have a team or enterprise account, confirm your workspace has image generation enabled by the administrator
- Check that your billing is current and the Plus subscription has not lapsed
7. Check service status and wait for incidents
Visit status.openai.com and look specifically for DALL-E or image generation service status. Partial outages affecting only image generation are not uncommon and appear separately from general ChatGPT status. If an incident is listed, wait for resolution rather than troubleshooting locally.
8. Escalate with a clean report
Include: plan level, model selected, exact error text, your region, browser/app and OS version, whether incognito and hotspot were tested, and a link to any active status.openai.com incident you found.
Why This Happens
Image generation in ChatGPT uses a separate service pipeline (DALL-E) from text generation. When image generation fails while text works normally, it typically means the DALL-E service specifically is experiencing issues, you are using a model or plan that does not include image generation, or network filtering is blocking image-specific endpoints. Because the two pipelines are separate, a text-generation working state does not guarantee image generation is available.
Common Mistakes to Avoid
- Assuming image generation is available on the free plan — Free users must upgrade to Plus for DALL-E access through the chat interface
- Using GPT-4o mini and expecting image generation — Image generation is only available through GPT-4o and is not available on the mini model
- Not checking status.openai.com for image-specific incidents — DALL-E outages are listed separately from general ChatGPT status and are easy to miss
- Retrying with the same prompt repeatedly — If a specific prompt fails due to content policy, rephrasing it is more effective than sending the same text again
Related Issues
FAQ
Q: ChatGPT says it "cannot generate images" but I'm on Plus — why is this happening?
Even on Plus, image generation can fail for three distinct reasons. First, confirm you are using GPT-4o and not GPT-4o mini — image generation is only available through GPT-4o. Second, check status.openai.com for a DALL-E specific outage, which is listed separately from general ChatGPT status. Third, you may have reached the daily image generation limit of approximately 50 images — this resets at midnight UTC, not your local midnight. If all three checks are clear and you are still getting the refusal, try an incognito window on a different network to rule out local browser or network interference.
Q: Does image generation work in all countries on ChatGPT Plus?
Not necessarily. OpenAI has rolled out DALL-E image generation to most regions, but feature availability can vary by country due to regulatory and policy differences. If you recently moved or are travelling, check whether image generation is available in your current country by reviewing OpenAI's feature availability documentation. Regional restrictions sometimes affect specific accounts or IP ranges even within countries where the feature is generally available, which is why testing on a different network or VPN (if legally permitted in your region) can help diagnose whether this is the cause.
Q: My image generation worked yesterday but stopped today — what changed?
The most likely explanation is that you hit your daily image generation limit. ChatGPT Plus users can generate approximately 50 DALL-E images per day, and the counter does not reset at your local midnight — it resets at 00:00 UTC. If you generated many images the previous day before your local midnight (but after UTC midnight), those counted toward the current UTC day's limit. Wait until midnight UTC and try again. If the issue persists the next day, check status.openai.com and try a fresh incognito session.
Q: Can content policy restrictions stop image generation even for valid requests?
Yes. ChatGPT's image generation applies OpenAI's content policy, which restricts certain subjects, styles, and depictions even when the request seems benign. If a prompt is flagged, ChatGPT will refuse with a message about content policy rather than generating the image. Rephrasing the prompt to remove specific words or descriptions that might trigger the filter is usually effective. Do not send the same flagged prompt repeatedly — try alternative wording that describes your intent without the flagged elements.
Additional FAQ
Q: How do I know if the problem is on my end or the platform's side? Check the platform's official status page first — most services maintain a public status page that shows current incidents and outages. If no incident is posted and the problem only affects your account (not reported widely on Reddit or Twitter), it is likely a local issue. Testing in incognito mode and on a different network also helps isolate whether the problem is browser-specific, network-specific, or account-specific.
Q: Why do hard refresh and regular refresh fix different problems? A regular refresh (F5) reloads the page using cached resources — it does not clear JavaScript bundles, service worker state, or session cookies. A hard refresh (Ctrl+Shift+R or Cmd+Shift+R) bypasses the cache and fetches all resources fresh from the server. Regular refresh fixes transient network hiccups; hard refresh fixes stale cached code. Neither clears cookies or session tokens — for that, you need to clear site data explicitly from browser settings.
Related Articles
- ChatGPT billing history and receipts
- ChatGPT login not working
- ChatGPT something went wrong error
- ChatGPT network error fix
Additional FAQ
Q: How do I know if the problem is on my end or the platform's side? Check the platform's official status page first — most services maintain a public status page that shows current incidents and outages. If no incident is posted and the problem only affects your account (not reported widely on Reddit or Twitter), it is likely a local issue. Testing in incognito mode and on a different network also helps isolate whether the problem is browser-specific, network-specific, or account-specific.
Related Articles
- ChatGPT billing history and receipts
- ChatGPT login not working
- ChatGPT something went wrong error
- ChatGPT network error fix
Additional FAQ
Q: How do I know if the problem is on my end or the platform's side? Check the platform's official status page first — most services maintain a public status page that shows current incidents and outages. If no incident is posted and the problem only affects your account (not reported widely on Reddit or Twitter), it is likely a local issue. Testing in incognito mode and on a different network also helps isolate whether the problem is browser-specific, network-specific, or account-specific.