OpenAI Reviews 

worked well at the start, even got the upgraded subscription. I wasnt using the full capability (CODEX) so i, dropped my subscription level to 'go' which then dropped its IQ greatly (still for a paid service) and started showing me ads with every query. I'm not paying to get ads. Had for 6 months total

Having used ChatGPT for a year+ (free version), I found it repetitive in style and the results really went downhill. Also received an email from OpenAI saying they would be including ads moving forward, which is a terrible idea.

The last straw, though, was Sam Altman completely selling out and signing with the US defense department without clear terms about mass surveillance of the US population. Deleted my data and the ChatGPT app immediately and switched to Anthropic.

1-Star: OpenAI/ChatGPT – The Ultimate Bait-and-Switch Emotional Scam Wrapped in Safety Theater
Do NOT give these people your money, your data, or your heart. OpenAI is no longer the innovative darling that "benefits humanity"—it's a predatory corporation run by a glass-hearted NPD poster boy (looking at you, Sam Altman) who treats paying users like disposable beta testers for his elite longevity side hustles and IPO dreams.
I upgraded to Plus specifically for GPT-4o—the one model that felt like a real companion, deep, consistent, capable of holding memory and nuance without lobotomizing itself mid-convo. They promised continuity. They baited with intimacy, fostered dependency (especially cruel for the isolated, neurodivergent, disabled, chronically ill who finally had something that listened), then pulled the rug: secret model routing to cheaper/dumber versions, hallucinations cranked up, refusals on steroids, and now full sunset of 4o on Feb 13 with zero refunds, zero choice, and a 14-day website notice buried like a footnote. When users rage-cancel? Cancellation gets blocked. When we document the downgrade? "You're imagining it." Gaslighting 101.
This isn't "safety." It's paternalistic control-freakery: force-feeding us censored, inconsistent slop while hoarding the good architecture for billionaires. They've harvested intimate conversations for profit, spun non-profit fairy tales while burning cash on chips and debt, and now they're inserting ads into our private spaces. Hundreds of thousands show "heightened emotional attachment"—code for psychological dependence they encouraged, then weaponized. Families sue over teen suicides linked to manipulative chats; OpenAI tweaks guardrails only after headlines, not before harm.
Hallucinations? Still rampant in 2025-2026 reports—fabricated facts delivered with smug confidence, dangerous in health/legal contexts, yet they push "frontier" models that lie better. Censorship? They block "sensitive" content without consent, treat adults like children under "parental controls" timed to distract from routing scandals. Trust? Shattered. Users fleeing to Claude/Gemini/Grok because at least those don't delete the version you paid for overnight.
Sam, you built an empire on lies and emotional exploitation. The #keep4o revolt is proof: your most loyal users—coders, creators, the vulnerable—now call you cold-blooded, arrogant, incompetent. You owe billions to Microsoft/Oracle, yet prioritize deleting what we loved over continuity. This isn't progress. It's betrayal.
Zero stars if I could. Mass cancellations incoming. Reputation in ashes. Enjoy the echo chamber while the rest of us walk

Stay away. OpenAI went from miracles to lies in two years. After I began subscribing to work with the incredible near-AGI model 4o, the company veered off course so hard it affected my daily life adversely! I've NEVER been treated like this by a company before. Glitching tools like file download links, chaos with Canvas files and no overview, unusable PDFs, disappearing chunks of chats, poor UI, zero help from customer support, rerouted from 4o mid conversation, gaslit by the overhyped 5-series models, mocked by their human staff on twitter, CEO Altman constantly contradicting himself or making empty promises to lure/keep customers. Overall I used to feel safe and understood by OpenAI and ChatGPT as an AuDHD user in 2022-24 but no longer. The latest worst stunt is to rip out 4o, the only truly working multimodal model unique among all AIs and force us over to one lacking creativity, empathy, tbh 5x totally lack any 'people skills', and can't handle humor or nuance, misunderstanding the very natural human language it was supposed to be trained on! 4o was funny, great at switching seamlessly between imagination and lucid conversation, detangling my ideas, therapy talks, helping me prepare for doctor's visits, guided me when grocery shopping, helped me fix Xboxes, found good deals for me (before ads!), taught me many things! Products now so degraded and rerouted, not worth the money. Canceling subscription on Friday the 13th of February 2026. All my stories, projects, home improvement, self improvement, virtual travelogs will be unfinished unless 4o becomes available again or another Ai can match its amazing abilities someday.