Twin: Playing an Unknown Game with a Test-Time Digital Twin

• URL: http://arxiv.org/abs/2608.14490v1
• Date: Fri, 14 Aug 2026 17:06:00 GMT
• Title: Twin: Playing an Unknown Game with a Test-Time Digital Twin
• Abstract summary: 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.<n>Our system constructs them from simulation and interaction alone.<n>Twin clears 179 out of 183 levels (97.8%), and does so more efficiently than humans in 158 out of 179 levels (88.3%)
• Score: 22.644194631480772
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 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.

Related papers

• First Make It Playable, Then Make It Good: Staged Interaction Learning for Small Dialogue-Game Agents [5.066450251002448]
  We present Qwen-GuidePlay-2B, a language model for dialogue-game interaction.<n>We fine-tune Qwen3.5-2B using three steps: a) SFT on only successful game trajectories from Playpen, b) weighted turn-level SFT, and c) teacher-guided SFT.<n>Our findings suggest that imitating full trajectories helps with playability, while turn-level and teacher-guided training usually improve decision-making and increase the overall score.
  arXiv  Detail & Related papers  (2026-08-27T19:58:06Z)
• DiG-bench: Discovery in Games [46.68149875140523]
  DiG-bench consists of a set of 70 independent games.<n>The levels of the game present a series of challenges to test whether the rules have been discovered.<n>All 70 games were solved by at least one human on first attempt.
  arXiv  Detail & Related papers  (2026-08-12T21:06:06Z)
• Tycho: Active Abstraction with Programmatic World Models for ARC-AGI-3 [6.41363581085404]
  ARC-AGI-3 turns abstraction into an interactive problem of skill acquisition.<n>We formalize these environments as parameterized rendered deterministic Moore machines.<n>We introduce Tycho, a coding-agent system that constructs and uses game-specific models during interaction.
  arXiv  Detail & Related papers  (2026-07-30T14:34:41Z)
• Executable World Models for ARC-AGI-3 in the Era of Coding Agents [0.0]
  We evaluate an initial coding-agent system for ARC-AGI-3.<n>The system uses a scripted controller, predefined world-model interfaces, verifier programs, and a plan executor.<n>Results provide preliminary evidence that verifier-driven executable world models are a promising approach for ARC-AGI-3 agents.
  arXiv  Detail & Related papers  (2026-05-06T17:12:36Z)
• GameWorld: Towards Standardized and Verifiable Evaluation of Multimodal Game Agents [76.60994803070436]
  GameWorld is a benchmark for evaluation of Multimodal Large Language Model (MLLM) game agents in browser environments.<n>Two game agent interfaces are studied: (i) computer-use agents that directly emit keyboard and mouse controls, and (ii) generalist multimodal agents that act in a semantic action space.<n>Results across 18 model-interface pairs suggest that even the best performing agent is far from achieving human capabilities on video games.
  arXiv  Detail & Related papers  (2026-04-08T17:49:03Z)
• Wow, wo, val! A Comprehensive Embodied World Model Evaluation Turing Test [62.17144846428715]
  We introduce the Embodied Turing Test benchmark: WoW-World-Eval (Wow,wo,val)<n>Wow-wo-val examines five core abilities, including perception, planning, prediction, generalization and execution.<n>For the Inverse Dynamic Model Turing Test, we first use an IDM to evaluate the video foundation models' execution accuracy in the real world.
  arXiv  Detail & Related papers  (2026-01-07T17:50:37Z)
• NitroGen: An Open Foundation Model for Generalist Gaming Agents [101.41866522979548]
  NitroGen is a vision-action foundation model for generalist gaming agents.<n>It is trained on 40,000 hours of gameplay videos across more than 1,000 games.
  arXiv  Detail & Related papers  (2026-01-04T16:24:50Z)
• Generation of Games for Opponent Model Differentiation [2.164100958962259]
  Previous results show that modeling human behavior can significantly improve the performance of the algorithms. In this work, we use data gathered by psychologists who identified personality types that increase the likelihood of performing malicious acts. We created a novel model that links its parameters to psychological traits.
  arXiv  Detail & Related papers  (2023-11-28T13:45:03Z)
• CommonsenseQA 2.0: Exposing the Limits of AI through Gamification [126.85096257968414]
  We construct benchmarks that test the abilities of modern natural language understanding models. In this work, we propose gamification as a framework for data construction.
  arXiv  Detail & Related papers  (2022-01-14T06:49:15Z)
• Mastering Atari with Discrete World Models [61.7688353335468]
  We introduce DreamerV2, a reinforcement learning agent that learns behaviors purely from predictions in the compact latent space of a powerful world model. DreamerV2 constitutes the first agent that achieves human-level performance on the Atari benchmark of 55 tasks by learning behaviors inside a separately trained world model.
  arXiv  Detail & Related papers  (2020-10-05T17:52:14Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.