Ernie Baidu's chatbot outperforms ChatGPT in several benchmark tests

Baidu has announced that the Ernie 3.5 artificial intelligence model behind its chatbot has outperformed OpenAI's ChatGPT in several key areas.

Here's What We Know

According to the tech giant, Ernie 3.5 outperformed GPT-3.5 in a number of benchmark tests. The model also outperformed the GPT-4 in Chinese language tests.

One of the tests was based on standardized entrance and qualifying exams. Ernie 3.5 outperformed both OpenAI models in Chinese. However, in English language tasks, the algorithm was inferior to GPT-4 but ahead of ChatGPT.

The test focuses on the Chinese language and contains over 13,000 multiple-choice questions covering over 50 different topics. Ernie 3.5 was ahead of ChatGPT and GPT-4 in this assessment.

The third test contains questions on various subjects in the fields of science and humanities. Ernie 3.5 is behind both ChatGPT and GPT-4.

Here's What We Know

Baidu has been publicly testing the Ernie bot in China since March 2023. It is based on the Ernie 3.5 language model. According to the developers, the chatbot can convert audio recordings into written notes, create emails, generate texts on prompting and more.

Source: CNBC.