Tencent’s latest release of GameGen-O, a diffusion transformation model for open-world video games, marks a new era of AI and game interaction. Different from traditional video generation models, GameGen-O can not only generate high-quality game content, but also allows users to control the actions of characters in the video in real time, bringing an unprecedented interactive experience. The training process is divided into two stages: pre-training and instruction fine-tuning. Through text-to-video and video continuation task learning, instruction-based video generation and interactive control are finally achieved. The breakthrough development of this technology brings new possibilities to both game developers and ordinary users.
Tencent recently launched an innovative video model called GameGen-O, which is the industry's first diffusion transformation model designed specifically for open-world video games. Different from traditional video models, GameGen-O not only generates high-quality game content, but also allows users to control the actions of characters in the video in real time like controlling game characters, creating a new era of artificial intelligence and game interaction.

The training process of GameGen-O is also unique, divided into two stages: basic model pre-training and instruction fine-tuning. In the pre-training stage, the model learns open-domain video game generation capabilities through text-to-video and video continuation tasks. In the instruction fine-tuning stage, the R&D team froze the pre-trained model and introduced trainable InstructNet for fine-tuning, so that the model can generate subsequent frames according to multi-modal structure instructions, thereby realizing instruction-based video generation and interactive control.

Although GameGen-O still has room for improvement in some aspects, it is undoubtedly an important milestone in AI-driven game content creation. This technology not only provides game developers with powerful tools, but also opens up a new era for ordinary users to freely create and explore virtual worlds. As technology continues to improve, we can expect that in the near future, everyone will be able to easily create their own immersive gaming experience.
The emergence of GameGen-O marks another in-depth integration of the game industry and the field of artificial intelligence. It not only demonstrates Tencent's strength in AI technology, but also points out the direction of future development for the entire industry. We wait and see how this technology will change the landscape of game creation and bring more surprises and possibilities to players.
Project address: https://top.aibase.com/tool/gamegen-o
The emergence of GameGen-O heralds the huge potential of AI technology in the gaming field. Its convenient interaction methods and powerful content generation capabilities will greatly promote the development of the gaming industry. In the future, GameGen-O may become a standard tool for game development, bringing players a more colorful gaming experience. We look forward to Tencent continuing to improve and bring us more surprises.