OpenAI has launched GPT-5.4 this week, enhancing the capabilities of ChatGPT for power users and professionals. This release follows the recent introduction of GPT-5.3 Instant, designed for everyday conversations. GPT-5.4 is touted as the “most capable and efficient frontier model for professional work,” targeting developers and industrial applications. Key improvements include enhanced efficiency, with GPT-5.4 now being the most token-efficient reasoning model to date, resulting in reduced token usage and faster problem-solving. It excels in real-world tasks, achieving a 83% success rate on GDPval tests, significantly outpacing its predecessor. The model also demonstrates a 33% decrease in false claims and an 18% drop in errors. A new flexible approach allows users to course-correct responses in real-time. Additionally, enterprise users can leverage the new ChatGPT for Excel add-in for sophisticated spreadsheet analysis. To access GPT-5.4, a paid ChatGPT subscription is required. For further details, visit OpenAI’s blog.
Source link
