Baidu, a Chinese tech giant, recently announced that their newest artificial intelligence (AI) model, Ernie 3.5, outperforms ChatGPT on several benchmark tests. In a report by China Science Daily, Ernie 3.5 surpassed ChatGPT and GPT-4 in two AI benchmarks, AGIEval, and C-Eval.
Baidu claims that Ernie 3.5 offers improved efficiency in both training and inference, resulting in faster and more cost-effective iterations in the future. Additionally, they stated that their new model would be compatible with external plugins and is more efficient than ChatGPT and GPT-4 on Chinese language tests, college qualifications, or bar exams for lawyers.
In a press release, Baidu announced, We are thrilled to announce the iteration of our foundation model, ERNIE, to version 3.5. ERNIE 3.5 has made significant strides in beta testing, surpassing ChatGPT (3.5) in comprehensive ability scores and outperforming GPT-4 in several Chinese language capabilities.
The initial release of the Ernie bot took place in Beijing back in March 2023, but it received a lukewarm response from the public, perceiving it as a rushed and underwhelming launch. However, the recent announcement of Ernie 3.5 may change the public’s perception of Baidu’s AI technology.
In contrast, OpenAI’s GPT-4 made waves in the AI community earlier this year with its multi-modal capabilities. Users can now input not only text but also pictures or voices, prompting the bot to perform specific tasks based on their input.
While the competition between Baidu and ChatGPT is undeniable, the advancements in AI technology will benefit society as a whole. AI advancements will continue to improve human lives by making work more efficient and reducing human error.
