Twin 让智能体边玩边写世界模型，ARC-AGI-3 得分从 7.8% 提到 93.3%，效率还超过人类。看看它怎么猜规则。
Twin 系统让前沿编程智能体在测试时编写可执行世界模型，用来完成 ARC-AGI-3 等持续学习任务，全程只靠模拟与交互推断游戏规则和目标。它在 183 个关卡中通过 179 个，成功率 97.8%，其中 158 个关卡的效率高于人类玩家（88.3%）。系统在 156 个已通关关卡中于任何奖励前就推断出目标（87.2%），其余关卡则靠搜索自动发现目标。直接用基础模型玩得分仅 7.8%，加现成 harness 升至 61.1%，而 Twin 世界模型把同一基础模型提升到 93.3%，打通 25 个游戏中的 23 个。
Twin: Playing an Unknown Game with a Test-Time Digital Twin
We present a Test-time World-model Inference (Twin) system, in which a frontier coding agent writes an executable world model for completing continual learning tasks, such as ARC-AGI-3 games. Traditional approaches hand-engineer such models, one custom design per task. Each game hides its rules and goal, and our system constructs them from simulation and interaction alone. Its inductive prior over grid games is strong enough to recover the true transitions of the game and the goal on nearly all levels. Replay validation happens in a twin world model. The harness enforces that an action is not made until the program reproduces every previous observed game transition. Each mismatch between a world model prediction and the actual action result becomes a counterexample that is used to repair the world model. Twin clears 179 out of 183 levels (97.8%), and does so more efficiently than humans in 158 out of 179 levels (88.3%). The system infers the goal before any reward on 156 of the levels it clears (87.2%), and in the remaining levels automatically discovers the goal by search. The benchmark scores completion and action efficiency, between 0 and 100, against humans playing each game for the first time. Played directly, the base model scores only 7.8%; an off-the-shelf harness increases it to 61.1%, whereas our twin world model increases the same base model to 93.3%, clearing 23 out of 25 games. Building a usable world model is simpler than anticipated, whereas the harder problem is inferring the right goal.