Last week, OpenAI launched GPT-5, but it didn’t quite live up to expectations. Fans who tried it found its responses short, less engaging, and full of silly mistakes. Many were especially frustrated because they couldn’t access the older GPT-4 model anymore, which they preferred. Instead, they got GPT-5, and it didn’t feel like a major step forward, despite what OpenAI’s CEO said.

What Went Wrong With GPT-5

GPT-5 was supposed to be a big upgrade, but the reality was different. It’s actually made up of two different models working together—a lightweight one for simple questions and a heavier one for more complex tasks. A special router model decides which one to use based on what’s asked. However, this router didn’t work well at launch and broke down, making GPT-5 seem dumber than it should have been. OpenAI has fixed it now, but many users still feel annoyed because they lost the choice of which model to use.

Cost-Cutting Measures or Genuine Innovation?

Another reason for the disappointment might be money. Some experts believe GPT-5 isn’t really about making a smarter AI but saving costs. To keep things affordable, OpenAI replaced its more capable models with this auto-switching system. This switch increases efficiency but reduces the quality of responses. Fans noticed that OpenAI now limits free users to just ten messages an hour and keeps the model’s memory size the same as before. Many in the community think GPT-5 is more about cutting expenses than real progress.

The Industry’s Bigger Picture

OpenAI is under pressure. It’s facing more competition and needs to turn a profit soon. The company is valued at around $500 billion, which means it’s expected to make money quickly. But fans have become attached to the quirks of the older models, even if they weren’t perfect. This has made it harder for OpenAI to fully shift to cost-saving measures without upsetting its users. The situation shows how tough it is to balance innovation, cost, and user satisfaction in AI development.

Despite the setbacks, OpenAI is still trying to improve GPT-5 and regain user trust. Many see this as a sign that the company is more focused on managing costs than pushing the boundaries of AI. For now, users will have to decide whether they want a cheaper, less capable AI or wait for future updates that might bring better performance. The debate about whether AI companies should prioritize innovation or profits is more relevant than ever.

Inspired by

• https://futurism.com/theory-why-gpt-5-sucks

Sources

• gmpg.org
• theguardian.com
• theregister.com
• techcrunch.com
• openai.com