OpenAI has launched GPT-5, a new standard model for all users. Those on the free tier would get limited usage access. ChatGPT “Plus” subscribers have higher usage limits. There is another expensive version, “Pro”, which costs $200 per month.
The Pro-tier offers more advanced features than the last two subscription models. It gives unlimited access to the more advanced reasoning model, GPT5 Thinking Pro. This model activates a deeper thinking mode when there is a complex query.
OpenAI has phased out older models, GPT-4o and GPT-4.1. Now, GPT-5 has variants with different costs and better performance, including GPT-5-mini and GPT-5-nano.
Microsoft has announced the integration of GPT-5 into the Copilot AI assistant.
Now there is no need for the manual model selection because GPT-5 operates as a unified system. It uses a “router” to decide whether to use a fast and efficient model or a robust thinking model for queries.
Sam Altman, CEO of OpenAI, said that GPT-5 possess an expert-level performance with significant improvements in writing, coding, math, and health-related queries.
“GPT-5 is the first time that it really feels like talking to an expert in any topic, like a PhD-level expert,” Altman said.
With extra coding capabilities, GPT-5 is described as “state-of-the-art on coding benchmark” that can create websites, apps, and it can write high-quality code. Altman “thinks this idea of software-on-demand is going to be one of the defining features of the GPT-5 era.”
The factual errors have dropped to 4.8% which is a direct result of the model’s significantly reduced hallucination rate.
The agentic capabilities have also been enhanced, which allows GPT-5 to use multiple tools to handle complex and multi-step tasks autonomously.
GPT-5 integrated all OpenAI tools in one interface, like image generation, voice capabilities, and web search. It can also get access to the user’s Google account and manage services like Google Calendar and Gmail.
A new personality feature allows users to select and adjust the Chatbot tone to Nerd, Cynic, Listener, and Robot. The multimodality of the model can process and generate text, image, and an updated voice mode that can adjust to pace and tone.
Sam Altman described GPt-5 as “a significant step along our path to artificial general intelligence (AGI).”
GPT-5 is a unified system that eliminates the need for manual model selection. It has improved reasoning, fewer factual errors and enhanced agentic capabilities. It can also handle multi-step tasks autonomously.