We are guided by what unites people
News
Articles
Media
Announcement
Technology
About Us
Contacts
||
OpenAI has launched its new language model, GPT-5, which fundamentally changes the approach to working with artificial intelligence. The main breakthrough lies in its ability to not just generate code, but to create complete projects from scratch. The model can develop finished web pages with interactive elements and styles based on a single prompt.
Among the key improvements is a significant increase in accuracy. The system has become better at debugging errors by analyzing interface screenshots. Additionally, it has learned to automatically select the optimal mode for processing a request: a quick mode for simple tasks and a deep analysis mode for complex ones, eliminating the need for the user to manually change settings. GPT-5 can also handle context up to one million tokens and is significantly less prone to inaccuracies and fabricating facts.
The gradual rollout of the new model began on August 7th, and it is now available to all user categories. The free version offers basic capabilities, while the Plus and Pro subscriptions provide extended functionality. The cost of access via API is lower than that of competitors and depends on the complexity of the task.
ORIENT
