# Reasoning on LiveBench (Acc, Expl., Fmt) Source: [Reasoning on LiveBench (Acc, Expl., Fmt)]() Collection index: [llms.txt]() Data updated: 2026-04-20T06:55:36.003Z Evaluation of model performance on reasoning tasks using the LiveBench dataset, measured by Accuracy, Explanation Score, and Format Score. 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 | Reasoning | | Dataset (including any stated split) | LiveBench | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | 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. - [Explanation Score](): Higher is better; unit: PERCENT. - [Format Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | LACE: + SFT & RL | 33 | Not recorded | 69e5b30c7d0172140ec7dfbc | | Independent Sampling: + Voting | 28 | Not recorded | 69e5b30c7d0172140ec7dfbc | | Independent Sampling: Standard SFT + RL | 25.6 | Not recorded | 69e5b30c7d0172140ec7dfbc | | LACE: + SFT & RL | 16.5 | Not recorded | 69e5b30c7d0172140ec7dfbc | | Independent Sampling: + Voting | 16 | Not recorded | 69e5b30c7d0172140ec7dfbc | | LACE: + SFT | 16 | Not recorded | 69e5b30c7d0172140ec7dfbc | | Isolated Parallel: + Voting | 15.5 | Not recorded | 69e5b30c7d0172140ec7dfbc | | Isolated Parallel: + Voting | 15.5 | Not recorded | 69e5b30c7d0172140ec7dfbc | | Independent Sampling: Standard SFT + RL | 14.6 | Not recorded | 69e5b30c7d0172140ec7dfbc | | Isolated Parallel: + SFT & RL | 14 | Not recorded | 69e5b30c7d0172140ec7dfbc | | Isolated Parallel: + SFT & RL | 11 | Not recorded | 69e5b30c7d0172140ec7dfbc | | LACE: + SFT | 7.5 | Not recorded | 69e5b30c7d0172140ec7dfbc | | Isolated Parallel: + SFT | 7 | Not recorded | 69e5b30c7d0172140ec7dfbc | | LACE: Continuous Pre-train | 5 | Not recorded | 69e5b30c7d0172140ec7dfbc | | Isolated Parallel: + SFT | 3.5 | Not recorded | 69e5b30c7d0172140ec7dfbc | | LACE: Continuous Pre-train | 1 | Not recorded | 69e5b30c7d0172140ec7dfbc | ### LACE: + SFT & RL Accuracy: 33. Result ID: 69e5cdca47f34ce924cb73ba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-4B"}. This record does not assert independent reproduction. ### Independent Sampling: + Voting Accuracy: 28. Result ID: 69e5cdca47f34ce924cb73b4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-4B"}. This record does not assert independent reproduction. ### Independent Sampling: Standard SFT + RL Accuracy: 25.6. Result ID: 69e5cdca47f34ce924cb73b3. [Reporting paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-4B"}. This record does not assert independent reproduction. ### LACE: + SFT & RL Accuracy: 16.5. Result ID: 69e5cdca47f34ce924cb73b2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Independent Sampling: + Voting Accuracy: 16. Result ID: 69e5cdca47f34ce924cb73ac. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### LACE: + SFT Accuracy: 16. Result ID: 69e5cdca47f34ce924cb73b9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-4B"}. This record does not assert independent reproduction. ### Isolated Parallel: + Voting Accuracy: 15.5. Result ID: 69e5cdca47f34ce924cb73af. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Isolated Parallel: + Voting Accuracy: 15.5. Result ID: 69e5cdca47f34ce924cb73b7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-4B"}. This record does not assert independent reproduction. ### Independent Sampling: Standard SFT + RL Accuracy: 14.6. Result ID: 69e5cdca47f34ce924cb73ab. [Reporting paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Isolated Parallel: + SFT & RL Accuracy: 14. Result ID: 69e5cdca47f34ce924cb73ae. [Reporting paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Isolated Parallel: + SFT & RL Accuracy: 11. Result ID: 69e5cdca47f34ce924cb73b6. [Reporting paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-4B"}. This record does not assert independent reproduction. ### LACE: + SFT Accuracy: 7.5. Result ID: 69e5cdca47f34ce924cb73b1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Isolated Parallel: + SFT Accuracy: 7. Result ID: 69e5cdca47f34ce924cb73b5. [Reporting paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-4B"}. This record does not assert independent reproduction. ### LACE: Continuous Pre-train Accuracy: 5. Result ID: 69e5cdca47f34ce924cb73b0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Isolated Parallel: + SFT Accuracy: 3.5. Result ID: 69e5cdca47f34ce924cb73ad. [Reporting paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### LACE: Continuous Pre-train Accuracy: 1. Result ID: 69e5cdca47f34ce924cb73b8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Scale":"Qwen3-4B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.