OpenAI introduces GPT-5.3 Instant: A new ChatGPT model saves users time
OpenAI has begun rolling out a new version of its language model — GPT-5.3 Instant. The main direction of the update is a fundamental change in the “behavior” of artificial intelligence: interaction with ChatGPT is expected to become more productive and less formal.
Key improvements and new features
Developers focused on the tone of communication and the quality of responses. The main changes include:
• Direct answers without long disclaimers: The model now refuses harmless questions less often and rarely uses lengthy introductory caveats.
• Deep integration with the internet: Web search capabilities have been improved. The neural network better combines fresh online data with its core knowledge to provide structured and up-to-date information.
• Accuracy and context: GPT-5.3 Instant better understands the nuance of user queries, maintains consistency during long conversations, and significantly reduces factual errors (hallucinations).
• Consistent tone: Within a single chat, the model maintains the chosen communication style, making it a more convenient tool for creative and professional tasks.
At present, GPT-5.3 Instant is already available to users with paid subscriptions. The previous version, GPT-5.2 Instant, remains temporarily in the Legacy Models section, but its support will be completely discontinued on June 3, 2026.
ORIENT
