OpenAI Launches the Sixth Version of ChatGPT

OpenAI began on Thursday to roll out the sixth version of ChatGPT, its most advanced artificial intelligence model, which can operate a computer independently, accompanied by a mechanism to curb it if it goes beyond its intended scope.

GPT-6 Astra was released a year after the previous version, following a series of cyberattacks this summer launched by AI tools under testing, including technologies developed by OpenAI and its competitor Anthropic, which revealed the risks posed by highly advanced models.

The sixth version will initially be available to a “limited number of organizations” before being rolled out “in the coming days” to subscribers of paid models and developers.

In a demo video, the company showcased its model, which is controlled by voice commands and independently publishes a sales advertisement. According to research results published by OpenAI, the model can find housing within 10 minutes, compared to six hours for a specialized employee.

GPT-6 is the first OpenAI model to reach the highest risk level according to the cybersecurity assessment scale. It can hack highly protected information systems without assistance.

Its most sensitive functions are restricted to cybersecurity specialists, and automated oversight can halt user tasks “including in operations unrelated to cybersecurity,” according to the company.