# Zero-shot Language Understanding on Qwen3-0.6B Evaluation Suite (test) Source: [Zero-shot Language Understanding on Qwen3-0.6B Evaluation Suite (test)]() Collection index: [llms.txt]() Data updated: 2026-05-19T11:38:18.261Z Evaluation of zero-shot performance across the Qwen3-0.6B specific suite, covering tasks such as AE, AC, SciQ, MM, PIQA, RACE, and others. 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 | Zero-shot Language Understanding | | Dataset (including any stated split) | Qwen3-0.6B Zero-shot Evaluation Suite (AE, AC, SciQ, MM, MM-P, HS, OBQA, PIQA, RACE, WG, CSQA, AGI) (test) | | Metric | Accuracy (AE) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 15 of 15 | | 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 (AE)](): Higher is better; unit: PERCENT. - [Accuracy (AC)](): Higher is better; unit: PERCENT. - [Accuracy (SciQ)](): Higher is better; unit: PERCENT. - [Accuracy (MM)](): Higher is better; unit: PERCENT. - [Accuracy (MM-P)](): Higher is better; unit: PERCENT. - [Accuracy (HS)](): Higher is better; unit: PERCENT. - [Accuracy (OBQA)](): Higher is better; unit: PERCENT. - [Accuracy (PIQA)](): Higher is better; unit: PERCENT. - [Accuracy (RACE)](): Higher is better; unit: PERCENT. - [Accuracy (WG)](): Higher is better; unit: PERCENT. - [Accuracy (CSQA)](): Higher is better; unit: PERCENT. - [Accuracy (AGI)](): Higher is better; unit: PERCENT. - [Average Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy (AE) | Result date | Report paper | | --- | --- | --- | --- | | AMO | 68.64 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | AMO | 68.56 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | AMO | 68.39 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | Muon-PE | 68.32 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | Muon-PE | 68.3 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | Muon-KJ | 67.47 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | Muon-KJ | 66.88 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | Muon-PE | 66.51 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | Muon-KJ | 66.5 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | Muon-PE | 66 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | AMO | 65.53 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | Muon-KJ | 65.32 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | Muon-PE | 65.17 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | Muon-KJ | 64.94 | Not recorded | 6a0beef0dbc836b87d82c8e0 | | AMO | 63.72 | Not recorded | 6a0beef0dbc836b87d82c8e0 | ### AMO Accuracy (AE): 68.64. Result ID: 6a0c4ba6fbc373592c036cfd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Optimizer":"AMO","Step":"6000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### AMO Accuracy (AE): 68.56. Result ID: 6a0c4ba6fbc373592c036d03. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Optimizer":"AMO","Step":"8580","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### AMO Accuracy (AE): 68.39. Result ID: 6a0c4ba6fbc373592c036d00. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Optimizer":"AMO","Step":"8000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### Muon-PE Accuracy (AE): 68.32. Result ID: 6a0c4ba6fbc373592c036cfc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Optimizer":"Muon-PE","Step":"6000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### Muon-PE Accuracy (AE): 68.3. Result ID: 6a0c4ba6fbc373592c036cff. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Optimizer":"Muon-PE","Step":"8000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### Muon-KJ Accuracy (AE): 67.47. Result ID: 6a0c4ba6fbc373592c036cfb. [Reporting paper]() Recorded evaluation configuration: {"Optimizer":"Muon-KJ","Step":"6000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### Muon-KJ Accuracy (AE): 66.88. Result ID: 6a0c4ba6fbc373592c036cfe. [Reporting paper]() Recorded evaluation configuration: {"Optimizer":"Muon-KJ","Step":"8000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### Muon-PE Accuracy (AE): 66.51. Result ID: 6a0c4ba6fbc373592c036cf6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Optimizer":"Muon-PE","Step":"2000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### Muon-KJ Accuracy (AE): 66.5. Result ID: 6a0c4ba6fbc373592c036d01. [Reporting paper]() Recorded evaluation configuration: {"Optimizer":"Muon-KJ","Step":"8580","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### Muon-PE Accuracy (AE): 66. Result ID: 6a0c4ba6fbc373592c036d02. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Optimizer":"Muon-PE","Step":"8580","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### AMO Accuracy (AE): 65.53. Result ID: 6a0c4ba6fbc373592c036cfa. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Optimizer":"AMO","Step":"4000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### Muon-KJ Accuracy (AE): 65.32. Result ID: 6a0c4ba6fbc373592c036cf8. [Reporting paper]() Recorded evaluation configuration: {"Optimizer":"Muon-KJ","Step":"4000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### Muon-PE Accuracy (AE): 65.17. Result ID: 6a0c4ba6fbc373592c036cf9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Optimizer":"Muon-PE","Step":"4000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### Muon-KJ Accuracy (AE): 64.94. Result ID: 6a0c4ba6fbc373592c036cf5. [Reporting paper]() Recorded evaluation configuration: {"Optimizer":"Muon-KJ","Step":"2000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. ### AMO Accuracy (AE): 63.72. Result ID: 6a0c4ba6fbc373592c036cf7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Optimizer":"AMO","Step":"2000","Backbone":"Qwen3-0.6B","Pre-training tokens":"60B + 36B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.