AnalysisAI ModelsNovember 14, 2026

Gamer: a vision-language-action foundation model for real-time game agents

Read original source →arxiv.org

Gamer combines perception, memory management and multi-step reasoning in one end-to-end agent framework for long-horizon tasks in 2D and 3D environments. Training uses continued pre-training, sparse-thinking instruction tuning and multi-turn reinforcement learning, plus a cloud-edge inference engine for unbounded-context execution.

3 sources

More stories today

Open the live feed