Share This Article
OpenAI has unveiled the next iteration of its language model for ChatGPT, GPT-4. The upgraded system now has the capability to accept images as inputs, although the output remains text-based. In a press release, OpenAI refers to the launch of GPT-4 as “the latest milestone in its effort in scaling up deep learning”.
The company boasts that GPT-4 has shown human-level performance on various professional and academic benchmarks.
For example, it passed a simulated bar exam with a score in the top 10% of test takers. However, despite its improved accuracy, the system is still prone to factual errors. OpenAI provides an example where GPT-4 referred to Elvis Presley as the son of an actor.
Despite its limitations, OpenAI has improved many of GPT-4’s safety properties. The company decreased the model’s tendency to respond to requests for disallowed content by 82% compared to GPT-3.5. And it responds to sensitive requests such as medical advice and self-harm per the company’s policies 29% more often.
“We look forward to GPT-4 becoming a valuable tool in improving people’s lives by powering many applications“, OpenAI wrote. “There’s still a lot of work to do, and we look forward to improving this model through the collective efforts of the community building on top of, exploring, and contributing to the model“.
OpenAI has released GPT-4’s text input capability into ChatGPT and its API, but only to a waitlist. However, the upgraded system is immediately available to paying users with the ChatGPT Plus $20 monthly subscription plan. People can also experience ChatGPT and GPT-4 in Microsoft Bing and the Edge browser, which the company announced in February.
- Google Brings AI Tools to Workspace and Opens Up PaLM for Developer
- Apple to Reconsider AI Development Strategy Amid Growing Competition
OpenAI has been rebuilding its deep learning stack over the past two years, and GPT-3.5 was the first test run of the new system. With the release of GPT-4, OpenAI hopes to continue improving the model through community contributions. Although the system is not yet perfect, the launch of GPT-4 represents an important step forward in the development of language models.