OpenAI recently released GPT-5, its newest AI model that promises better reasoning and improved coding skills. CEO Sam Altman said it’s the best AI the company has ever built. But early reactions from users have been underwhelming, raising questions about whether the industry is truly making progress or just spending more money for less.

Many power users expected GPT-5 to be a big step forward. Instead, they found it to be a mixed bag. Some say it produces shorter answers that aren’t any better than previous models. Others complain about restrictions that limit how much they can use it. A popular post on Reddit called GPT-5 “horrible,” criticizing its short replies, less personality, and the fact that users hit usage limits quickly. Many fear that OpenAI will face backlash once the excitement around the release fades.

Adding to the controversy is OpenAI’s decision to deprecate all older models. That means users can no longer access the versions they relied on for their work. This move has upset many, especially those who preferred the older, more flexible tools. It’s a risky gamble, especially since the AI industry is investing huge sums to develop these models. Critics wonder if OpenAI’s latest release is really the best it can do, or if cost-cutting is behind the scenes.

Rumors about GPT-5 have been circulating for over a year and a half. Many expected a revolutionary upgrade, but instead, it feels more like a series of minor improvements and setbacks. Some think OpenAI might be trying to save money by making models more efficient, which could explain the lack of noticeable progress. Running large AI models is energy-intensive and environmentally costly, so cost savings could be a factor.

Many users see GPT-5 as a disappointment wrapped in a strong brand. Reddit comments describe answers as shorter and less useful, and the restrictions as worse than before. Some joke about mourning their old AI “friends,” feeling that the new model feels abrupt or overworked. The official GPT-5 system card, which outlines its capabilities, didn’t impress many either. Critics pointed out that it doesn’t show clear improvements on many coding benchmarks.

However, there might be a silver lining. A nonprofit organization called METR, which studies the risks of advanced AI, found that GPT-5 is unlikely to speed up AI research by more than ten times or be used maliciously. This suggests that, despite its shortcomings, GPT-5 isn’t a threat to society or a breakthrough in AI power. Still, OpenAI’s CEO has not directly addressed the negative feedback yet.

Altman did acknowledge that GPT-5 is the “smartest model” they’ve made, but he emphasized that the focus was on making AI accessible and affordable for billions of people. He promised that more advanced versions are still on the way, and that these will benefit many. Given OpenAI’s high valuation, the company appears eager to keep the hype alive, even if the current release leaves some users feeling let down.

In short, GPT-5’s launch has sparked a lot of debate. While OpenAI claims it’s a step toward widespread AI adoption, many early adopters are disappointed, questioning whether the model lives up to the hype. As the AI industry continues to pour money into development, only time will tell if these new models will truly deliver on their promises or if cost-cutting is holding back genuine progress.

Inspired by

• https://futurism.com/gpt-5-sucks

Sources

• gmpg.org
• openai.com
• openai.com