# Puzzle Solving on Sokoban (Out Of Distribution) Source: [Puzzle Solving on Sokoban (Out Of Distribution)]() Collection index: [llms.txt]() Data updated: 2026-03-23T11:44:27.146Z Evaluation of model performance on solving Sokoban puzzles using the Out Of Distribution (OOD) test set, measured by Avg@128 and Pass@128. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Puzzle Solving | | Dataset (including any stated split) | Sokoban (Out Of Distribution) | | Metric | Avg@128 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 12 of 12 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Avg@128](): Higher is better; unit: PERCENT. - [Pass@128](): Higher is better; unit: PERCENT. ## Recorded results | Method | Avg@128 | Result date | Report paper | | --- | --- | --- | --- | | Markov | 66.9 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | Markov | 31.6 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | State-action-sequence | 30.2 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | State-action-sequence | 20.4 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | State-action-sequence | 0.2 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | Action-sequence | 0 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | Action-sequence | 0 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | Markov | 0 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | Action-sequence | 0 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | Action-sequence | 0 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | Markov | 0 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | | State-action-sequence | 0 | Not recorded | 69c0c8ed0636e5bf4afc3aaa | ### Markov Avg@128: 66.9. Result ID: 69c1272247f34ce924a9b409. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-3B-It","Training Stage":"RL post-training"}. This record does not assert independent reproduction. ### Markov Avg@128: 31.6. Result ID: 69c1272247f34ce924a9b403. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen3-4B","Training Stage":"RL post-training"}. This record does not assert independent reproduction. ### State-action-sequence Avg@128: 30.2. Result ID: 69c1272247f34ce924a9b405. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen3-4B","Training Stage":"RL post-training"}. This record does not assert independent reproduction. ### State-action-sequence Avg@128: 20.4. Result ID: 69c1272247f34ce924a9b40b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-3B-It","Training Stage":"RL post-training"}. This record does not assert independent reproduction. ### State-action-sequence Avg@128: 0.2. Result ID: 69c1272247f34ce924a9b404. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen3-4B","Training Stage":"SFT Warm-up"}. This record does not assert independent reproduction. ### Action-sequence Avg@128: 0. Result ID: 69c1272247f34ce924a9b400. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen3-4B","Training Stage":"SFT Warm-up"}. This record does not assert independent reproduction. ### Action-sequence Avg@128: 0. Result ID: 69c1272247f34ce924a9b401. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen3-4B","Training Stage":"RL post-training"}. This record does not assert independent reproduction. ### Markov Avg@128: 0. Result ID: 69c1272247f34ce924a9b402. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen3-4B","Training Stage":"SFT Warm-up"}. This record does not assert independent reproduction. ### Action-sequence Avg@128: 0. Result ID: 69c1272247f34ce924a9b406. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-3B-It","Training Stage":"SFT Warm-up"}. This record does not assert independent reproduction. ### Action-sequence Avg@128: 0. Result ID: 69c1272247f34ce924a9b407. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-3B-It","Training Stage":"RL post-training"}. This record does not assert independent reproduction. ### Markov Avg@128: 0. Result ID: 69c1272247f34ce924a9b408. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-3B-It","Training Stage":"SFT Warm-up"}. This record does not assert independent reproduction. ### State-action-sequence Avg@128: 0. Result ID: 69c1272247f34ce924a9b40a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen2.5-3B-It","Training Stage":"SFT Warm-up"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.