The General Langauge Understanding Evaluation (GLUE) is the gold-standard of AI language benchmarking.

Baidu Artificial Intelligence has overtaken two of the world’s leaders in the industry. The Chinese tech giant topped Microsoft and Google in a recent artificial intelligence competition.

The contest centered around how well a machine can understand human languages. Baidu artificial intelligence received the highest score ever in the General Language Understanding Evaluation (Glue).

Baidu Artificial Intelligence Surpasses Human Achievement

Its Ernie model – short for Enhanced Representation through kNowledge IntEgration – achieved the first-ever score above 90. It also easily beat U.S. firms and university labs. It is also one of 10 AI systems to ever achieve the average human score of 87.1.

The company’s model topped Google’s Bert (short for Bidirectional Encoder Representations from Transformers). Both models are named after Sesame Street characters.

According to The Independent – the firm’s platform surpassed its rivals due in part to its design process. Company developers first programmed Ernie by using the Chinese language. It then developed Ernie to understand English. The report also states that the developers realized that this process only made it better at understanding the English language.

“When we first started this work, we were thinking specifically about certain characteristics of the Chinese language. But we quickly discovered that it was applicable beyond that,” Hao Tian, Baidu Research’s top architect, recently told MIT Technology Review.

Related: Artificial Intelligence: Walmart’s Cutting Edge Use of AI and Machine Learning in Retail