OpenAI’s GPT-5 reportedly falling short of expectations

OpenAI's development of its next major model, GPT-5, is experiencing delays and high costs, as reported by The Wall Street Journal. The model, code-named Orion, has undergone at least two large training runs over an 18-month period, but the results have not yet justified the expenses involved. The initial training run was slower than anticipated, indicating that further training would be both time-consuming and costly. Although GPT-5 shows improvements over previous models, it has not advanced enough to warrant the cost of maintaining it.

In addition to using publicly available data and licensing deals, OpenAI has hired individuals to generate new data by writing code and solving math problems. The company is also utilizing synthetic data created by another model, o1. OpenAI has not provided a comment on the situation, and it has previously stated that it would not release a model named Orion this year.

Key takeaways:

```html

OpenAI's development of GPT-5, code-named Orion, is behind schedule and not yet cost-effective according to a report by The Wall Street Journal.
The development of GPT-5 has included at least two large training runs, but the initial run was slower than expected, indicating high time and cost requirements.
OpenAI is employing new strategies, including hiring people to create fresh data and using synthetic data from another model, to enhance GPT-5's capabilities.
The company has stated that it will not release a model code-named Orion this year.

```

OpenAI’s GPT-5 reportedly falling short of expectations | TechCrunch

Key takeaways:

Comments (0)

Newsletter