Show HN: Jevman – AI decision models play Pac-Man

Category: ai-ml

Tags: ai-benchmark, pac-man, decision-models, leaderboard, typescript

Score: 6.8/10 (Innovation: 7, Technical: 7, Documentation: 8, Utility: 5)

JevMan is a benchmark and interactive playground that evaluates AI decision models by having them play Pac-Man in real time against classic arcade ghosts. It measures model performance through a leaderboard and allows users to add their own models, serving as a practical testbed for comparing decision-making AI.

Target audience: AI researchers, machine learning engineers, and developers evaluating decision-making models

Repository: https://opper.ai/jevman-benchmark/ · TypeScript · AGPL-3.0 · 20 stars

View on Hacker News