What models built
One page per model: passed pages first, then failed pages from recent runs. Each runs in a sandbox with no network and no access to this site.
Minesweeper
-
Qwen3.8-27B-FastShows whether a model keeps the rules of a small game exact: where an empty area stops spreading, what a flag does to the count, and what a win or a loss reveals.6k tokens · M4 Max 128 GB · uwu · passed