China’s Baidu is looking to reclaim lost ground with two new AI models
Can the newly-released models actually beat rivals like DeepSeek or OpenAI's ChatGPT?
The AI race in China is getting spicier by the minute. Just days after Alibaba dropped an AI model that can supposedly read emotions, Baidu entered the chat with two new models of its own—ERNIE 4.5 and ERNIE X1.
ERNIE X1, according to Baidu, matches DeepSeek’s R1 in performance—at half the price. But Baidu isn’t stopping at just affordability. X1 is also being pitched as the first “deep-thinking” AI that can reason, plan, and even improve itself over time. In other words, it’s not just answering questions—it’s trying to think like a human.
Then there’s ERNIE 4.5, Baidu’s latest multimodal model, which means it can process and generate text, images, video, and audio instead of just text. And because AI can’t just be smart, it apparently has a personality too—Baidu says ERNIE 4.5 has “high Emotional Intelligence (EQ)” and can understand memes, satire, and internet humour.
Kelechi Edeh
Now, one major reason we’ve been seeing China’s tech giants—Alibaba, Tencent, and now Baidu—pushing out newer and more affordable AI models could be DeepSeek’s unexpected rise earlier this year. But for Baidu, there might be another angle at play here.
While the company was one of China’s first to introduce a ChatGPT-style AI, its Ernie Bot never really took off. Now, it’s back with a reasoning-focused AI (X1) and an upgraded multimodal model (4.5), likely in a bid to reclaim lost ground.
But of course, promises are one thing, real-world performance is another. While Baidu’s models sound great on paper, can they actually beat DeepSeek or OpenAI's ChatGPT where it matters? Or is this just another AI model in China's crowded ~$80 billion AI market?
Acquah Nana Yeboah
