Giant Network Group is focusing on Sora's breakthrough progress and actively laying out the multi-modal large model field.

On February 20th, Giant Network Group (002558.SZ) replied to investors on the Shenzhen Stock Exchange's interactive platform, stating that the company has been paying attention to the breakthrough progress made by Sora in the field of video generation. As one of the earliest game companies to complete the filing of large language models, Giant Network Group is also actively laying out in the field of multimodal large models, focusing on the creation scene of game sub-categories. In the short term, Giant Network Group believes that Sora has the potential to greatly improve the efficiency of content creation in the gaming industry as a productivity tool. In the long term, the understanding of 3D models and the strong ability for physical world simulation emerging from Sora will have far-reaching impacts on the underlying tools of games. The company is actively laying out in the field of multimodal large models, focusing on the creation scene of game sub-categories, and conducting in-depth research and exploration on the stability and controllability of AI video generation and 3D model generation algorithms. The company has widely applied its self-developed AI video style transfer technology in the research and operation process to empower content creativity. It is worth mentioning that Giant Network Group has previously completed the filing of its game AI large model, GiantGPT, making it the first enterprise in the gaming industry to publicly confirm this achievement, marking an important progress for the company in the field of artificial intelligence. GiantGPT is a large model specialized in the gaming industry, trained with high-quality proprietary data and internet public data, and deeply optimized for basic abilities such as character interpretation, scene reasoning, and long-term memory. Its performance in Giant's own gaming business scenarios significantly surpasses closed-source general large models both domestically and internationally. Furthermore, the company's AIGC production capabilities have been preliminarily established at the corporate level, with the construction of platform-based basic technological capabilities such as "iMagine - AI painting cloud platform", "CodeBrain - code generation assistant", and "BaiLing - AI voice generation", covering directions in visual, speech, large language models, etc., to enhance efficiency in research, operation, and distribution processes.