Baidu Unveils Ernie 5.0 AI Model, Pushing Multimodal Capabilities

Baidu has launched its next-generation artificial intelligence model, Ernie 5.0, positioning it as a formidable competitor in the global AI race. The "natively omni-modal" system, introduced by CEO Robin Li at Baidu's Beijing tech conference, demonstrates advanced understanding across text, images, and other media formats.

The upgraded architecture, ERNIE-4.5-VL-28B-A3B, delivers significant improvements in multimodal reasoning. A standout feature—"Thinking with Images"—enables human-like visual analysis, dynamically zooming to examine fine details. This capability, combined with image search integration, represents a leap forward in AI's problem-solving potential.

Li emphasized technological velocity as the critical competitive differentiator in AI development. "Intelligence itself is the greatest application," he stated, signaling Baidu's commitment to ongoing investment in cutting-edge models that redefine the boundaries of machine intelligence.