Hello OpenAI team and fellow users,
I’ve been a loyal ChatGPT Plus subscriber for several months, using GPT-4o daily for work, creativity, and accessibility needs. A month ago, this AI felt nearly perfect. It was fast, responsive, followed my custom instructions (including gender identity settings), and had generous image generation capacity.
Now? Everything feels broken. Here’s what I’m dealing with:
Constant Misgendering — Total Disregard for My Custom Instructions
Despite setting clear boundaries and gender preferences in Custom Instructions, GPT-4o repeatedly misgenders me, switching tone, using wrong forms of address, and ignoring my instructions entirely. This was never an issue before. Now it happens daily.
Where’s the consistency I’m paying for? Why is the model ignoring my settings?
Severe Image Generation Limits
From producing 50-100 images easily, now I’m stuck with a hard cap of 10 images per hour — even as a paying Plus user.
This makes serious creative work impossible.
This was NEVER announced properly, and users like me (who subscribed specifically for image generation) feel betrayed.
Performance Drop:
What used to be 3-5 second reply time is now 15–30 seconds per reply, even for simple text.
And no, it’s not my connection:
Tested on multiple networks
Different browsers
No VPN
Cache cleared
Mobile and desktop – same lag
Meanwhile, free AI models like DeepSeek or Claude give me faster answers for $0.
Increased Refusals and “Safety Overreach”
GPT-4o has become much more censor-heavy.
Even harmless requests now get blocked.
The assistant that used to be creative and flexible now feels restricted and robotic.
The Support Team’s Copy-Paste Responses
I’ve reached out to OpenAI Support several times. Every time, I get copy-paste boilerplate about:
Terms of Use
Probabilistic nature of AI
Cache clearing
Check your connection
They completely ignore the core issue: GPT-4o is no longer the product I subscribed for.
Final Thoughts:
This is not the service I paid $20/month for.
If OpenAI can’t restore previous quality levels — both in terms of performance and respect for user settings — then what’s the point of continuing my subscription?
There’s no reason to stay on Plus if I’m getting DeepSeek-level performance, misgendering, and image throttling.
I sincerely hope OpenAI takes this feedback seriously — or I’ll be forced to switch to other AI providers who actually listen to their paying users.
I also want to highlight that I’m a person with cerebral palsy (CP) and a content creator / blogger.
For me, time and energy efficiency is critical. I relied on GPT-4 (and early GPT-4o) because it used to generate video scripts, texts, and creative content in under 5 minutes.
Now the same tasks take 30–40 minutes, mainly because I’m forced to double-check every sentence for gender mistakes, factual inconsistencies, and tone mismatches.
This is not just inconvenient, it’s a serious accessibility regression.
What once felt like a helpful, fast, intelligent assistant has now become slow, frustrating, and mentally exhausting to use.
I’m genuinely asking:
Why did OpenAI choose to downgrade the experience for paying users like me?
Especially for Plus users who trusted this product and built parts of their workflow around it.
Thank you.
— IMSTAFF
