Video game design is becoming one of the clearest test cases for 3D-native artificial intelligence. Tencent’s Hunyuan family of models can generate 3D objects and interactive scenes, and that capability is already being explored inside major game development workflows.
According to a researcher familiar with the effort who spoke on the condition of anonymity, developers at Riot Games, a Tencent subsidiary, are using 3D-native AI models to prototype new characters, scenes, and storylines for Valorant. The same source says Tencent’s models are also being used by developers of another Tencent game, GKART, and by some independent developers. Tencent declined to comment.
Why 3D AI Fits Game Development
Many widely discussed AI systems generate text, images, or video. Hunyuan is aimed at a different part of the creative pipeline: the objects and environments that make up interactive worlds.
That distinction matters for video games. Games are not only viewed; they are navigated, tested, revised, and played. A character concept, object, or scene has to work as part of a space, not just as a flat image.
The source described the shift in practical terms: “The games industry requires a lot of investment,” the source says. “Previously you would need a month to design a character. Now you can just type in some text, and Hunyuan can give you four choices in 60 seconds.”
That does not mean a finished game character appears without human judgment. But it does suggest a faster early stage, where artists and designers can test more possibilities before committing to a direction.
What Hunyuan Can Generate
Tencent’s Hunyuan, written as 混元 or “first mix,” includes models that can produce 3D objects and interactive scenes. HunyuanWorld 1.0, released in July, generates interactive scenes. A more basic model, Hunyuan 3D, can create 3D objects.
In October, Tencent released a new version of HunyuanWorld that lets users upload video to generate 3D scenes. That points to a broader ambition: using different kinds of input to produce interactive spatial output.
The WIRED source article describes testing HunyuanWorld 1.0 in a scene that looked like part of a Lego movie, with a valley of brightly colored blocks stretching into the distance. It also describes using Hunyuan 3D to make custom Dungeons & Dragons characters for 3D printing.
For game teams, the important idea is not only that 3D content can be generated. It is that the generated material can match the form of assets game developers already work with. Alexander Raistrick, a graduate student at Princeton University working on novel approaches to generating 3D content, says video games are an obvious use case because “Outputting 3D meshes [a standard way of representing 3D objects] is your typical kind of bread and butter of game development.”
A Broader Race Around 3D Worlds
Hunyuan is part of a larger turn in AI research toward models that can work with the physical world, or with digital spaces that behave more like it. Many experts believe that future AI systems will need a deeper understanding of that world in order to advance.
Tencent is not alone. Microsoft, Meta, Stability AI, and Bytedance all offer 3D models. Hunyuan sits at the top of one leaderboard designed to rank such tools, according to the source article.
Startups and academic projects are also pushing into the same area. World Labs, founded by Fei-Fei Li, has developed Marble, a tool that produces fully consistent and persistent 3D scenes. The source article notes that this could help generate games on the fly or create reliable training data for robots.
At Stanford University, a project called 3D Generalist used an LLM to decide how to modify scenes with new objects. Raistrick is developing a code-based approach to generating 3D scenes, which could let LLMs generate and interact with scenes in a more powerful way. Google DeepMind’s SIMA 2 shows how AI agents could interact with virtual worlds and create new forms of gameplay.
The Promise Comes With Tension
The use of AI in game creation remains controversial. As in other creative fields, concerns about AI-fueled job loss are significant. Some developers argue that games should be labeled when they contain AI-made content, while others say the technology is already widespread in the industry.
The debate is likely to sharpen as 3D AI becomes more useful. Text and image generation can influence concept work and marketing material. But 3D-native models move closer to the actual building blocks of interactive entertainment: characters, props, scenes, and playable spaces.
That raises practical and creative questions for studios:
- How much of early prototyping should be automated?
- Where should human artists and designers remain central in the process?
- Should players be told when AI-generated content appears in a game?
- How will development teams judge quality when many more options can be created quickly?
The source article does not resolve those questions. It does show why they are becoming harder to avoid.
Why Tencent May Have an Edge
Tencent’s position is notable because the company already has deep experience with games and digital platforms. The source article notes that Tencent produces some of the world’s most popular video games and movies, operates WeChat, and has its own chatbot, YuanBao, integrated into WeChat.
Those assets do not guarantee leadership in 3D AI. But they give Tencent a strong reason to keep developing models that can build and understand interactive worlds. If 3D-capable AI becomes more central to game design, Tencent’s existing game development experience may become an advantage.
For now, the clearest signal is inside the development process itself. Hunyuan is being tested as a tool for faster prototyping, not as a full replacement for game creators. But the direction is clear: the next wave of AI in games is not only about writing dialogue or generating images. It is about making worlds that designers can enter, adjust, and eventually turn into playable experiences.