OpenAI launches GPT-5 to curb AI hallucinations

Artificial intelligence users and developers alike have long wrestled with a common issue: AI “hallucinations,” where models confidently generate false or misleading information. Addressing this challenge head-on, OpenAI has announced GPT-5 — their most advanced AI system yet — designed to dramatically reduce hallucinations and deliver expert-level accuracy across a broad range of tasks.

Source: www.unsplash.com

The new model boasts state-of-the-art capabilities in coding, mathematics, writing, healthcare, visual perception, and more. What sets GPT-5 apart is its unified architecture, which dynamically chooses when to respond quickly and when to engage in deeper reasoning to tackle complex queries, leading to more reliable and nuanced answers.

Dr Chris Blair, group director of 21st Century

AI: The new frontier in SA's psycho-legal assessments - part 1

Chris Blair 8 Aug 2025

At the heart of GPT-5’s innovation is a real-time routing system that intelligently directs requests to either a fast-response model or a more powerful “thinking” model, depending on the difficulty of the question or the user’s instructions. This router continuously learns from user interactions, improving its ability to deliver the most accurate response possible.

Beyond reducing hallucinations, GPT-5 also improves instruction-following and curbs sycophantic tendencies, making it more practical for real-world applications. Writing quality is elevated, coding capabilities reach new heights — particularly in generating sophisticated front-end designs and debugging large projects — and health-related answers show greater precision and care.