Baidu ERNIE multimodal AI beats GPT and Gemini in benchmarks

Baidu's latest multimodal AI model, ERNIE-4.5-VL-28B-A3B-Thinking, has achieved notable success by surpassing both GPT and Gemini in critical benchmarks. This advancement is particularly important as it focuses on enterprise data types that are frequently neglected by conventional text-centric models. By effectively analyzing complex data sources like engineering schematics, factory-floor video feeds, medical scans, and logistics dashboards, Baidu's ERNIE model promises to unlock valuable insights for businesses. The ability to interpret such diverse data is crucial for organizations seeking to enhance operational efficiency and make informed decisions. As AI continues to evolve, the implications of this development could lead to a significant shift in how enterprises utilize AI technologies, emphasizing the need for models that can handle multimodal data effectively.