Recently, a significant development has taken place as the Institute of Automation at the Chinese Academy of Sciences, in collaboration with the Wuhan Institute of Artificial Intelligence, has officially launched the Zidong Taichu 4.0 multimodal reasoning large - scale model.

Since its inaugural release in 2021, this model has gone through four successive iterations. Over this period, it has achieved a remarkable technological leap. Initially, it was limited to pure text - based thinking and simple image - assisted thinking. However, it has now evolved into a system capable of fine - grained multimodal semantic thinking. This transformation marks the dawn of a new era in multimodal deep reasoning.

Zidong Taichu 4.0 is endowed with proactive thinking capabilities. These capabilities empower it to "observe, recognize, and reason simultaneously." In terms of performance, it outperforms GPT - 5 in several key areas. These include long - video understanding, image - assisted reasoning, and multimodal task processing.

The model has already found practical applications, having been deployed in more than 60 vertical scenarios across various fields. For instance, it is being utilized in intelligent welding and medical diagnosis, demonstrating its versatility and potential for real - world impact.

To complement the model, the "Zidong Taichu Cloud" platform has been introduced. This platform offers a comprehensive suite of full - stack services. These services span from computing power support and model training to application development, providing a one - stop solution for users.

Moreover, the model has forged partnerships with 28 ecological collaborators. Through these collaborations, an industrial ecosystem has been built. To further solidify its presence, the model's national operations headquarters has been established in Wuhan Optics Valley.