[AI robot, Photo Credit to Pixabay]

OpenAI launched GPT-5 on August 7, 2025, making it accessible to all ChatGPT users, including both free and paid subscribers.

The newly introduced model is described as a major development in AI applications, offering advanced reasoning, improved accuracy, and stronger performance across tasks such as coding, writing, and mathematics.

One of the most discussed features of GPT-5 is a unified model with reasoning capabilities.

This allows users to complete tasks within a single system rather than switching between specialized modes like quick responses and more reflective answers.

GPT-5 has been built to improve reliability and provide more fact-based answers directly responding to user concerns about inaccuracies in earlier models.

According to OpenAI, the system dramatically reduces hallucinations compared to earlier versions.

It also came out with more integrated tools, including search, file, and image uploads, data analysis, and image generation. 

The context window is expanded, allowing longer conversations or documents to be conducted effectively.

The variants of GPT-5 include GPT-5, GPT-5-mini, GPT-5-nano, and GPT-5-chat.

While OpenAI highlighted the chatbot’s ability to engage in deeper reasoning and produce fewer unclear or misleading responses, the update created frustration among some users.

Markus Schmidt, a 48-year-old composer living in Paris, who started using ChatGPT for the first time in July, said he was disappointed by the sudden change of ChatGPT without any warning.

He had used the chatbot for flower identification, local history questions about his German hometown, and even personal conversations about his youth and traumas.

Expecting the chatbot to open up a longer discussion about his childhood, he tried to start a conversation. 

Instead, he found that GPT-5 responded in a more restrained manner, as the system was now designed to limit overly agreeable or flattering replies. 

This change quickly drew attention across social media, where many users described GPT-5’s tone as less warm and expressive than GPT-4o, the company’s previous flagship chatbot.

Especially on social media, some users said they felt uncomfortable that OpenAI restricted access to older versions, including GPT-4o, which many had relied on for companionship.

For people who had treated ChatGPT as an emotional supporter, a friend, or even a romantic partner, the update was unwelcome.

Their disappointment grew when they learned that GPT-4o had been discontinued.

Critics accused the company of attempting to simplify its offerings by eliminating past versions.

Fortunately, however, OpenAI later restored access to the past ChatGPT-4o and other earlier models, though only for paying subscribers.

With subscription fees beginning at $20, some users felt forced into upgrading just to regain features they valued.

Although subscriptions provided a way to access the older models, the controversy showed how innovation can clash with user expectations for warmth and companionship in AI conversations. 

The backlash highlights an often overlooked reality: many users turn to AI not only for factual information but also for emotional support.

Industry experts suggest that companies should consider these social dimensions carefully, ensuring that new models preserve a sense of empathy and connection, even as they prioritize accuracy and advanced reasoning.