V-GameGym: Visual Game Generation for Code Large Language Models

Wei Zhang, Jack Yang, Renshuai Tao, Lingzheng Chai, Shawn Guo, Jiajun Wu, Xiaoming Chen, Ganqu Cui, Ning Ding, Xander Xu, Hu Wei, Bowen Zhou · 2025

Read V-GameGym: Visual Game Generation for Code Large Language Models by Wei Zhang, Jack Yang, Renshuai Tao, Lingzheng Chai, Shawn Guo, Jiajun Wu, Xiaoming Chen, Ganqu Cui, Ning Ding, Xander Xu, Hu Wei, Bowen Zhou (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-20136.txt.

Description

Wei Zhang and eleven co-authors present a benchmark for code-generating language models in this 2025 arXiv paper. Current coding tests mostly check syntax and correct execution. Real game development also demands playability, good looks, and engaging interaction. V-GameGym was built to measure that fuller set of skills.

The benchmark holds more than two thousand samples grouped into one hundred themes drawn from real code repositories. A clustering method kept the set diverse and structurally complete. Grading uses an automated multimodal pipeline inside full interface sandboxes. Models must synthesize visual game code, not only plain functions. The analysis shows the benchmark separates plain coding skill from practical game-building ability. It gives developers measurable targets for visual programming quality.

Catalog record

CREATORWei Zhang, Jack Yang, Renshuai Tao, Lingzheng Chai, Shawn Guo, Jiajun Wu, Xiaoming Chen, Ganqu Cui, Ning Ding, Xander Xu, Hu Wei, Bowen Zhou
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE61 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-20136

Browse the full library · About the Tower of Babel