MiHoYo team open-sources GameScaling models that control games through keyboard and mouse actions

A research team involving miHoYo’s Honkai AI researchers and university collaborators has open-sourced GameScaling, a family of models that play games by generating keyboard and mouse actions from visual observations. The release includes eight checkpoints ranging from 800 million to 27 billion parameters, built through continued pretraining of Qwen3.5 vision-language models. The models use current and recent game frames, together with previous actions, to produce short sequences of controls. Demonstrated tasks include mining and crafting in Minecraft, combat and navigation in Doom, and puzzle-solving in escape-room environments. The project provides model weights, inference code and example programs under the Apache 2.0 license, allowing developers to study how training across different virtual worlds affects interactive AI performance. [GameScaling]