OpenAI Launches GPT-5, Now Available for Free and Paid Users
OpenAI has officially released GPT-5, its most advanced AI model yet, just days after announcing its first open-weight models in six years. The new model is now accessible to all users, including those on free accounts, marking a significant upgrade in performance, safety, and usability.
During a press briefing, OpenAI CEO Sam Altman compared the leap from previous models to GPT-5 as akin to moving from old pixelated screens to Retina Displays. He emphasized that GPT-5 feels like conversing with an expert in any field, offering a unified and more intuitive experience. Unlike earlier versions, GPT-5 is now the default reasoning model, eliminating the need for users to manually select advanced options like Think Longer.
GPT-5 comes in three variants. Free and Plus users get access to the standard version, with GPT-5 Mini stepping in when usage limits are reached. Pro subscribers, paying $200 per month, can use GPT-5 Pro without restrictions. Plus users enjoy significantly higher query limits than free users, while Pro customers have unlimited access.
One of GPT-5’s standout improvements is its reasoning speed. Altman noted that the model responds so quickly that users might question whether it’s thinking deeply enough—until they see the quality of its answers. OpenAI also claims GPT-5 reduces hallucinations by approximately 80% compared to its predecessor, o3, making it more reliable for factual accuracy.
Safety has also been enhanced with a new feature called Safe Completions. Instead of a binary approach where prompts are either accepted or rejected, GPT-5 now provides the most helpful answer within safety constraints. For sensitive topics, such as potentially harmful queries, the model offers high-level information without enabling misuse.
Health-related queries see improvements as well, with GPT-5 better at flagging concerns and suggesting when to consult a healthcare provider. It also tailors responses based on the user’s knowledge level and location.
Beyond reasoning and safety, GPT-5 excels in coding and writing, producing more compelling and precise outputs. OpenAI has also introduced new ChatGPT features, including customizable chat colors and deeper Google integrations for Gmail, Calendar, and Contacts. Pro users will get these Google features first, with wider availability later.
The Custom Instructions panel now includes preset personalities—cynic, robot, listener, and nerd—allowing users to tweak how ChatGPT interacts with them. Additionally, OpenAI has retired its Standard Voice Mode, replacing it with an upgraded Advanced Voice Mode that better adapts to speech nuances.
When asked about GPT-5’s role in achieving artificial general intelligence (AGI), Altman expressed skepticism about the term but acknowledged the model’s significant leap in capability. While GPT-5 still can’t learn continuously on its own, its intelligence level represents a major step forward.
With GPT-5, OpenAI continues pushing the boundaries of AI, delivering a faster, safer, and more capable model accessible to all users. The rollout includes new features and refinements, reinforcing OpenAI’s position at the forefront of AI innovation.
