AnalysisAI ModelsOctober 9, 2026

Jevman benchmark pits AI decision models against Pac-Man

Read original source →opper.ai

Six decision models each played 100 games against the classic scripted ghosts, with a 2-second deadline per junction decision and games capped at 5 minutes. The longest game lasted 2 minutes 24 seconds; rankings use mean score with a 95% margin of error.

1 source

More stories today

Open the live feed