Reddit users are not holding back on their feelings about the recent GPT-5 update. The general sentiment? Confusion, frustration, and disbelief. The new release promised smarter responses, advanced voice options, and an overall leap forward — but many feel the reality falls short.
User Feedback and Frustrations
One user summed it up bluntly: the update feels like a ‘gaslighting mess,’ with GPT-4o slowly becoming less capable while GPT-5 introduces inconsistencies and underwhelming features. Features that were once helpful now stumble, context is easily lost, and creative output reportedly suffers.
Users have tested the models across coding projects, creative writing, and other tasks. The verdict appears consistent: GPT-5 struggles where GPT-4o used to shine. Projects hang, memory fails, and the new Advanced Voice Mode is widely criticized. Meanwhile, the Reddit community debates whether this is a cost-saving measure, a subtle push toward subscription tiers, or simply the growing pains of increasingly complex models.
High-Value Community Comments
- AstralHippies: “It can’t code like it used to, it can’t do creative writing like it used to, can’t follow instructions like it used to, has no personality.”
- JaxLikesSnax: While benchmarks suggest improvements, real-world functionality is lacking and lobotomization is evident: “Instead of big drama and fake benchmarks I would rather wait longer for actual ingenuity and an honest product.”
- Gloomy-Detail-7129: Notes that the once comforting and collaborative nature of GPT-4o is now frustratingly absent, leaving users exhausted.
Takeaways
The core issue seems to be a mismatch between marketed improvements and the day-to-day experience of users, leaving a vocal portion of the community feeling unheard. While some adapt or fine-tune their experiences, many long for the reliability of previous versions.
Read the full discussion on Reddit here.
Leave a Reply