OpenAI has released its latest artificial intelligence (AI) model, GPT-5.4, in response to a reported decline in user numbers and increasing competition from companies like “Anthropic.” The company positions GPT-5.4 as its most capable and efficient model to date for professional applications.
The recent model is available in three versions: standard, “Pro” – designed for demanding, high-performance tasks – and “Thinking,” which is geared towards problem-solving and analysis. OpenAI emphasizes that GPT-5.4 operates more efficiently than previous iterations, accomplishing the same tasks using fewer tokens, resulting in faster processing and reduced costs.
According to company statements, GPT-5.4 has demonstrated strong performance in various tests, including operate with computer programming, information analysis, and complex task completion. The company highlights the model’s particular aptitude for time-consuming tasks requiring in-depth analysis, such as creating presentations, performing financial calculations, and analyzing legal documents.
OpenAI also reports a significant reduction in errors with the new model, noting that the generation of inaccurate or misleading information has decreased by approximately 30% compared to the previous version. This improvement in accuracy could be a key differentiator as the AI market matures and demands greater reliability. The release of GPT-5.4 underscores OpenAI’s commitment to remaining a leader in the rapidly evolving AI landscape, a sector attracting significant investment and scrutiny.
