OpenAI's ChatGPT-5 criticized

OpenAI has just launched GPT-5, promoting it as the smartest, fastest, and most useful artificial intelligence (AI) model ever.

However, the Reddit community found GPT-5 to be less than expected and even more limited than previous versions. A post titled “GPT-5 is horrible” received 4,600 upvotes and 1,700 comments, reflecting dissatisfaction. Users complained that GPT-5’s responses were short, lacked detail, and that its communication style was “too AI-like” and lacked personality.

ChatGPT Plus users are also limited to 200 messages per week and find it inconvenient to only use GPT-5 and not be able to revert to older models. Another reviewer compared GPT-5 to “having a severe brain injury,” referring to its inefficient responses.

Still, some users consider GPT-5 a significant upgrade over GPT-4, with many new features. "GPT-5 is clearly a big step forward," said AI editor Alex Hughes, but limitations such as message limits and the inability to access older models have disappointed many. Another user suggested that GPT-5's problems may be temporary and will improve.

According to Bloomberg, some of the negative reactions may stem from confusion about how the model works. Unlike previous versions, GPT-5 automatically switches the appropriate model based on a user’s query.

Source: https://nld.com.vn/chatgpt-5-cua-openai-bi-che-19625081221524752.htm