# Sokoban on Sokoban (Hard-2) Source: [Sokoban on Sokoban (Hard-2)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:30:49.177Z Evaluation of model performance on the Hard-2 difficulty setting of the Sokoban environment, measured by Accuracy. 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 | Sokoban | | Dataset (including any stated split) | Sokoban (Hard-2) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 14 of 14 | | 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 - [Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | GPT-5 | 87.89 | Not recorded | 692d2147d6b0dd3342931901 | | OpenAI o3 | 85.16 | Not recorded | 692d2147d6b0dd3342931901 | | OpenAI o4-mini | 62.89 | Not recorded | 692d2147d6b0dd3342931901 | | WMAct | 49.9 | Not recorded | 692d2147d6b0dd3342931901 | | PPO - Interactive | 46.83 | Not recorded | 692d2147d6b0dd3342931901 | | Claude 4.5 Sonnet | 29.3 | Not recorded | 692d2147d6b0dd3342931901 | | Gemini 2.5 Pro | 20.31 | Not recorded | 692d2147d6b0dd3342931901 | | Qwen3-14B | 5.08 | Not recorded | 692d2147d6b0dd3342931901 | | Qwen3-8B-Own | 1.39 | Not recorded | 692d2147d6b0dd3342931901 | | Qwen3-8B | 1.17 | Not recorded | 692d2147d6b0dd3342931901 | | PPO - EntirePlan | 0.35 | Not recorded | 692d2147d6b0dd3342931901 | | GPT-4o | 0 | Not recorded | 692d2147d6b0dd3342931901 | | Qwen2.5-7B-Instruct | 0 | Not recorded | 692d2147d6b0dd3342931901 | | Qwen2.5-32B-Instruct | 0 | Not recorded | 692d2147d6b0dd3342931901 | ### GPT-5 Accuracy: 87.89. Result ID: 699635875698665ec477e7a3. [Reporting paper]() Recorded evaluation configuration: {"category":"Proprietary Models"}. This record does not assert independent reproduction. ### OpenAI o3 Accuracy: 85.16. Result ID: 699635875698665ec477e7a4. [Reporting paper]() Recorded evaluation configuration: {"category":"Proprietary Models"}. This record does not assert independent reproduction. ### OpenAI o4-mini Accuracy: 62.89. Result ID: 699635875698665ec477e7a5. [Reporting paper]() Recorded evaluation configuration: {"category":"Proprietary Models"}. This record does not assert independent reproduction. ### WMAct Accuracy: 49.9. Result ID: 699635875698665ec477e7af. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### PPO - Interactive Accuracy: 46.83. Result ID: 699635875698665ec477e7ae. [Reporting paper]() Recorded evaluation configuration: {"mode":"interactive"}. This record does not assert independent reproduction. ### Claude 4.5 Sonnet Accuracy: 29.3. Result ID: 699635875698665ec477e7a7. [Reporting paper]() Recorded evaluation configuration: {"category":"Proprietary Models"}. This record does not assert independent reproduction. ### Gemini 2.5 Pro Accuracy: 20.31. Result ID: 699635875698665ec477e7a6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Proprietary Models"}. This record does not assert independent reproduction. ### Qwen3-14B Accuracy: 5.08. Result ID: 699635875698665ec477e7ab. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Opensource Models"}. This record does not assert independent reproduction. ### Qwen3-8B-Own Accuracy: 1.39. Result ID: 699635875698665ec477e7ac. [Reporting paper]() Recorded evaluation configuration: {"backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### Qwen3-8B Accuracy: 1.17. Result ID: 699635875698665ec477e7aa. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Opensource Models"}. This record does not assert independent reproduction. ### PPO - EntirePlan Accuracy: 0.35. Result ID: 699635875698665ec477e7ad. [Reporting paper]() Recorded evaluation configuration: {"mode":"single-turn output"}. This record does not assert independent reproduction. ### GPT-4o Accuracy: 0. Result ID: 699635875698665ec477e7a2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Proprietary Models"}. This record does not assert independent reproduction. ### Qwen2.5-7B-Instruct Accuracy: 0. Result ID: 699635875698665ec477e7a8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Opensource Models"}. This record does not assert independent reproduction. ### Qwen2.5-32B-Instruct Accuracy: 0. Result ID: 699635875698665ec477e7a9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Opensource Models"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.