According to Cointelegraph, Baidu, a leading Chinese technology company, has launched version 4.0 of its popular artificial intelligence (AI) large language model (LLM) chatbot, Ernie. CEO Robin Li claims that the updated model's overall performance is on par with GPT-4. The new Ernie model has been improved in four key areas: understanding, generation, reasoning, and memory.

In terms of understanding, Baidu reports significant enhancements in human-computer interaction. Ernie Bot can now accurately interpret out-of-order statements, vague expressions, and implied meanings in text. Notably, the model has been trained in both Chinese languages and English, which has traditionally been a challenge for LLM engineers.

Compared to previous versions, Ernie 4.0 has also made substantial progress in the quality and speed of generating images, videos, and coherent copy. In the reasoning category, CEO Robin Li demonstrated the model's advanced analytical problem-solving capabilities by posing a complex question. Unlike past LLM models, Ernie provided a correct answer, along with text and imagery.

Lastly, the 4.0 update expanded Ernie's memory capabilities. Although there is no scientific evidence that LLMs or AI systems can reason or have actual memory, the model's ability to process problems and recall prompts and outputs from previous sessions is comparable to ChatGPT. According to Robin Li, Ernie Bot can remember previously generated content even after five rounds of conversation and writing thousands of words.