OpenAI GPT-5 Development Falls Behind Schedule

According to reports, OpenAI's efforts to develop its next major AI model, GPT-5, have reportedly fallen behind schedule, with the results so far failing to justify the significant costs involved. This news aligns with earlier reports from The Information, which suggested that GPT-5 might not represent the same groundbreaking leap seen in previous iterations, prompting OpenAI to explore new strategies.

The report reveals more details about the 18-month development process of GPT-5, codenamed Orion. According to sources, OpenAI has completed at least two major training runs aimed at enhancing the model by training it on large datasets. However, the initial runs took longer than expected, indicating that larger-scale operations are both time-consuming and costly. While GPT-5 outperforms its predecessors, the progress made so far has not been sufficient to justify the high expenses required to maintain the model.

In addition to publicly available data and licensed agreements, OpenAI has employed individuals to create new data by writing code or solving mathematical problems. The company has also utilized synthetic data generated by another of its models, o1.

OpenAI has not provided an immediate comment on the situation. The company previously stated that it would not release the Orion model this year.