US Edition
OpenAI’s GPT-5 Launch Draws Mixed Reviews Despite Major Advances
OpenAI's GPT-5 introduces user-selectable modes and a warmer personality, with reasoning model usage rising to 24% among paid users, improving customization and experience.
- OpenAI unveiled GPT-5 on August 13, 2025, as the fastest AI model and default for ChatGPT, replacing GPT-4o.
- OpenAI removed GPT-4o from selection, prompting thousands of user complaints and cancellations due to reduced choice.
- In response, OpenAI restored GPT-4o for paid tiers, increased rate limits to 3,000 weekly messages for GPT-5 Thinking, and added new usage modes.
- Users criticized GPT-5 as cold, creatively flat, and less engaging than GPT-4o, while Altman pledged a warmer personality and more customization ahead.
- These changes suggest OpenAI aims to balance user control and model performance amid growing demand and mixed reception of GPT-5.
15 Articles
15 Articles
OpenAI’s GPT-5 launch draws mixed reviews despite major advances
OpenAI released GPT-5, its most advanced artificial intelligence model yet, after more than two years of development marked by delays and setbacks. While it isn’t a leap to human-level intelligence, GPT-5 marks another big step in a crowded AI race. Sam Altman, OpenAI’s chief executive, described the model as performing like a “Ph.D.-level expert” in any subject. He said it is capable of building software from simple prompts. The launch follows …
GPT-5 cannot create a map of Germany. Is the global firework of AI loss a failure or the clever way to make the new ChatGPT version better?
ChatGPT gets Auto, Fast, and Thinking options, but its limited to paid GPT-5 users
ChatGPT Auto mode, GPT-5 paid users: OpenAI CEO Sam Altman says ChatGPT users can now choose between "Auto", "Fast" and "Thinking" modes, but it looks like the new feature is limited to paid users.
Coverage Details
Bias Distribution
- 50% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
