We're launching our leaderboard for FrontierCode 1.1, a private evaluation from Cognition that tests models on completing coding tasks in a way that w...