NitroGen: An Open Foundation Model for Generalist Gaming Agents

TL;DR

NitroGen is a comprehensive vision-action foundation model trained on extensive gameplay videos, demonstrating strong cross-game generalization and transferring effectively to unseen games across various genres.

Contribution

Introduces NitroGen, a unified large-scale vision-action model trained on 40,000 hours of gameplay, with new datasets and benchmarks for evaluating cross-game generalization.

Findings

Achieves up to 52% improvement in task success rates on unseen games.

Demonstrates competence in diverse gaming domains including 3D combat, 2D platforming, and exploration.

Provides publicly available datasets, benchmarks, and model weights.

Abstract

We introduce NitroGen, a vision-action foundation model for generalist gaming agents that is trained on 40,000 hours of gameplay videos across more than 1,000 games. We incorporate three key ingredients: 1) an internet-scale video-action dataset constructed by automatically extracting player actions from publicly available gameplay videos, 2) a multi-game benchmark environment that can measure cross-game generalization, and 3) a unified vision-action model trained with large-scale behavior cloning. NitroGen exhibits strong competence across diverse domains, including combat encounters in 3D action games, high-precision control in 2D platformers, and exploration in procedurally generated worlds. It transfers effectively to unseen games, achieving up to 52% relative improvement in task success rates over models trained from scratch. We release the dataset, evaluation suite, and model weights to advance research on generalist embodied agents.