Tokyo AI (TAI) | OpenAI : 100-Hour Game Builder Challenge

Tokyo AI (TAI) runs a 100-hour build challenge with support from OpenAI, opening with a kickoff for selected teams on Friday, September 11 at 18:00 and starting the clock at 20:00 that evening. Teams begin from an empty repository and submit by 23:59 on Tuesday, September 15. Entrants pick one of two tracks: an AI-native game where generative AI drives characters, dialogue, rules or mechanics, or a tool that changes how games are developed, tested, localized, operated or distributed. Applications run from August 29 to September 10, or until the spots are filled. Solo entries are allowed but teams are encouraged, and OpenAI engineers offer optional technical support during the build window. Judging weighs meaningful use of OpenAI technology (30%), originality (25%), playability or utility (25%), and execution and craft (20%). Each submission needs a working demo, a short write-up, the chosen track, and an explanation of how OpenAI technology was used. The top 10 to 15 teams present at the final showcase at Shibuya Sakura Stage on Thursday, September 17 from 18:00 to 21:00, with all participants invited whether or not their team presents. Judges are Justin Waldron, Jerry Chi, Shunsuke Oyu, Rob Pereyda and Tyler Ryu. Prizes exceed US$100,000 in OpenAI API credits plus ChatGPT Pro subscriptions, with the winning team also invited to a matchmaking event with leaders from Japan's gaming industry. Applicants must be 18 or older, resident in a ChatGPT-supported country, and finalists must be able to present in person in Shibuya. Sessions run in both Japanese and English.