The world of artificial intelligence is about to take a major leap forward. OpenAI has confirmed that it will release GPT-5, the next-generation model behind ChatGPT, in early August. According to CEO Sam Altman, the update is set to dramatically expand reasoning, comprehension, and memory capabilities. 

One of the most anticipated improvements lies in memory management. While GPT-4o currently handles up to 128,000 tokens, GPT-5 will scale beyond one million.

This will allow the model to analyze entire databases in a single run, sustain context-rich conversations lasting weeks, and even process information accumulated over years. 

Chat GPT-5: an updated version

GPT-5 will arrive in three different versions: the standard model, a “mini” version, and a “nano” edition designed for lighter tasks or low-resource environments. With this tiered approach, OpenAI aims to serve both enterprise-level needs and individual users. 

Beyond being a natural evolution of GPT-4 turbo (launched in November 2023), the new release will also integrate reasoning capabilities from the parallel “o-series” models. This hybrid approach is intended to combine the broad knowledge base of GPT models with the logical and problem-solving strengths of the newer series. 

Availability will be rolled out gradually. Subscribers of ChatGPT Pro will be the first to gain access, followed by limited rollout to Plus users. Later, free-tier users may also get access, though with restrictions. 

With this launch, OpenAI reinforces its role as a global leader in artificial intelligence, at a time when smart automation is becoming an increasingly practical tool in daily life.