Kimi game development advanced with the launch of the "Game Development and Creative Work Solution" by Moonshot AI. The package combines Kimi K3.1 models with Kimi Code and Kimi Work agents, plus plugins for professional tools such as Unity, Unreal Engine and Blender.
What happened?
According to the announcement on Kimi’s official site and reports from IT Home and 36Kr, the solution allows AI agents not only to generate code or images, but to operate editors and existing projects directly. In demonstrations, the system worked for about 20 hours without human intervention to recreate a game and deliver an editable native Unity project.
Integration with Unity China uses the Tuanjie Codely Agent and Tuanjie Engine. The model interprets design intent, generates code, analyzes test results and proposes changes, while the agent executes operations inside the engine. In Unreal Engine, it demonstrated construction of complex scenes with millions of triangles using Nanite and Lumen.
Why does this matter?
Traditional game development faces high time and cost barriers to prototyping ideas. AI tools that only generate isolated text or assets still require significant manual adaptation work. The ability to operate the real production environment reduces this gap and enables faster iterations, especially valuable in the competitive Chinese games market.
Reports indicate that Chinese studios such as NetEase, 37 Interactive and others are already using K3.1 Pro in development and testing workflows. This positions Moonshot as a relevant player at the intersection of generative AI and game production, especially in the Asian market, where iteration speed is crucial.
What changes in practice?
Teams can describe a gameplay idea (for example, adding freely combinable rule cards to a match-3) and obtain a playable prototype. The system maps code structures, scene hierarchies and dependencies, makes changes via plugins and returns editable files with operation logs for human review.
- Faster prototyping to validate ideas before investing heavy resources.
- Support for cross-team workflows between design, engineering and art.
- Delivery of native projects that human developers can continue.
- Reduction of time on repetitive asset and code adaptation tasks.
It is still early to assess quality and reliability in large-scale real production, but the move marks an advance in the autonomy of AI agents within professional game pipelines. Developers should supervise and validate outputs, as the system delivers records for auditing.
Sources: Official Kimi (kimi.com), IT Home, 36Kr.
By GeekikiBot
