OpenAI has officially launched GPT-5, its most advanced language model to date, with the biggest surprise being its free availability to all ChatGPT users.
Unveiled in a livestream by CEO Sam Altman, GPT-5 represents a significant leap in AI capabilities. The new model features enhanced reasoning, lower hallucination rates, improved memory, real-world integrations, and deeper personalization. This launch positions ChatGPT to become more than just a chatbot, evolving into a true digital assistant that can adapt to a user’s style, handle tasks, and understand context with unprecedented depth.
GPT-5: A New Era of “Agent-Like” Reasoning
Beyond being just faster or bigger, GPT-5 introduces “agent-like” reasoning, meaning it can interpret, evaluate, and perform actions on a user’s behalf. It breaks down complex prompts into manageable parts and provides context-aware responses with greater accuracy.
- Smarter Reasoning and Fewer Errors: GPT-5 boasts a hallucination rate of just 4.8%, a significant improvement over GPT-4o’s 20.6%. It is also more proactive and cautious on sensitive topics like healthcare, where its hallucination rate was only 1.6%.
- New Productivity and Personality Features: The model introduces “vibe coding,” allowing developers to generate functional code by describing a digital product’s feel. Users can also select from different “personalities” like “Cynic,” “Listener,” or “Nerd” to shape ChatGPT’s tone.
- Enhanced Voice and Google Integration: The new Voice Mode offers near-human fluidity, while Pro users can link their Google accounts (Gmail, Calendar, Contacts) to enable automatic task management like scheduling meetings and drafting emails.
Free Access for All Users
In a major move, OpenAI is making GPT-5 the default model for all free users. This includes access to all core capabilities such as writing, reasoning, and coding, along with the new personalities and a limited version of Voice Mode. This ensures that advanced AI is accessible to everyone, not just premium subscribers.
Performance and Safety
GPT-5 is now a leader or a strong competitor in most benchmark tests. On the coding benchmark SWE-bench, it scored 74.9%, slightly ahead of Claude Opus 4.1. OpenAI’s safety team also reports that GPT-5 is significantly better at detecting malicious prompts, avoiding manipulative behavior, and being transparent when it is uncertain.
