# Mathematical Reasoning on Kaoyan (pass@1) Source: [Mathematical Reasoning on Kaoyan (pass@1)]() Collection index: [llms.txt]() Data updated: 2026-03-11T07:55:21.798Z Evaluation of model performance on mathematical reasoning tasks using the Kaoyan dataset, measured by pass@1 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 | Mathematical Reasoning | | Dataset (including any stated split) | Kaoyan | | Metric | pass@1 Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 19 of 19 | | 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 - [pass@1 Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | pass@1 Score | Result date | Report paper | | --- | --- | --- | --- | | SGPO | 73.37 | Not recorded | 68fa7bbf1e209131cea8a2ae | | GRPO | 71.36 | Not recorded | 68fa7bbf1e209131cea8a2ae | | Baseline | 64.32 | Not recorded | 68fa7bbf1e209131cea8a2ae | | SGPO | 58.29 | Not recorded | 68fa7bbf1e209131cea8a2ae | | SGPO | 57.79 | Not recorded | 68fa7bbf1e209131cea8a2ae | | SGPO | 56.78 | Not recorded | 68fa7bbf1e209131cea8a2ae | | GRPO | 55.78 | Not recorded | 68fa7bbf1e209131cea8a2ae | | SGPO | 54.77 | Not recorded | 68fa7bbf1e209131cea8a2ae | | SGPO | 52.26 | Not recorded | 68fa7bbf1e209131cea8a2ae | | Baseline | 50.25 | Not recorded | 68fa7bbf1e209131cea8a2ae | | GRPO | 48.24 | Not recorded | 68fa7bbf1e209131cea8a2ae | | SGPO | 48.24 | Not recorded | 68fa7bbf1e209131cea8a2ae | | Baseline | 45.73 | Not recorded | 68fa7bbf1e209131cea8a2ae | | GRPO | 43.22 | Not recorded | 68fa7bbf1e209131cea8a2ae | | SGPO | 39.7 | Not recorded | 68fa7bbf1e209131cea8a2ae | | SGPO | 38.69 | Not recorded | 68fa7bbf1e209131cea8a2ae | | Baseline | 37.69 | Not recorded | 68fa7bbf1e209131cea8a2ae | | GRPO | 35.68 | Not recorded | 68fa7bbf1e209131cea8a2ae | | Baseline | 29.15 | Not recorded | 68fa7bbf1e209131cea8a2ae | ### SGPO pass@1 Score: 73.37. Result ID: 69b11fe66dbb7075b95b4d92. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QwQ-32B","Judge Model":"DeepSeek-V3-0324"}. This record does not assert independent reproduction. ### GRPO pass@1 Score: 71.36. Result ID: 69b11fe66dbb7075b95b4d91. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"QwQ-32B"}. This record does not assert independent reproduction. ### Baseline pass@1 Score: 64.32. Result ID: 69b11fe66dbb7075b95b4d90. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"QwQ-32B"}. This record does not assert independent reproduction. ### SGPO pass@1 Score: 58.29. Result ID: 69b11fe66dbb7075b95b4d86. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B","Judge Model":"QwQ-32B","Stability Parameters":"w/o {beta,gamma}"}. This record does not assert independent reproduction. ### SGPO pass@1 Score: 57.79. Result ID: 69b11fe66dbb7075b95b4d82. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B","Judge Model":"o4-mini-0416"}. This record does not assert independent reproduction. ### SGPO pass@1 Score: 56.78. Result ID: 69b11fe66dbb7075b95b4d84. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B","Judge Model":"Qwen3-235B-A22B"}. This record does not assert independent reproduction. ### GRPO pass@1 Score: 55.78. Result ID: 69b11fe66dbb7075b95b4d81. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### SGPO pass@1 Score: 54.77. Result ID: 69b11fe66dbb7075b95b4d83. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B","Judge Model":"DeepSeek-V3-0324"}. This record does not assert independent reproduction. ### SGPO pass@1 Score: 52.26. Result ID: 69b11fe66dbb7075b95b4d85. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B","Judge Model":"QwQ-32B"}. This record does not assert independent reproduction. ### Baseline pass@1 Score: 50.25. Result ID: 69b11fe66dbb7075b95b4d80. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### GRPO pass@1 Score: 48.24. Result ID: 69b11fe66dbb7075b95b4d8e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-32B-Instruct"}. This record does not assert independent reproduction. ### SGPO pass@1 Score: 48.24. Result ID: 69b11fe66dbb7075b95b4d8f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-32B-Instruct","Judge Model":"o4-mini-0416"}. This record does not assert independent reproduction. ### Baseline pass@1 Score: 45.73. Result ID: 69b11fe66dbb7075b95b4d8d. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-32B-Instruct"}. This record does not assert independent reproduction. ### GRPO pass@1 Score: 43.22. Result ID: 69b11fe66dbb7075b95b4d8b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-14B-Instruct"}. This record does not assert independent reproduction. ### SGPO pass@1 Score: 39.7. Result ID: 69b11fe66dbb7075b95b4d89. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Llama-8B","Judge Model":"Claude-3.7"}. This record does not assert independent reproduction. ### SGPO pass@1 Score: 38.69. Result ID: 69b11fe66dbb7075b95b4d8c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-14B-Instruct","Judge Model":"o4-mini-0416"}. This record does not assert independent reproduction. ### Baseline pass@1 Score: 37.69. Result ID: 69b11fe66dbb7075b95b4d8a. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-14B-Instruct"}. This record does not assert independent reproduction. ### GRPO pass@1 Score: 35.68. Result ID: 69b11fe66dbb7075b95b4d88. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Llama-8B"}. This record does not assert independent reproduction. ### Baseline pass@1 Score: 29.15. Result ID: 69b11fe66dbb7075b95b4d87. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Llama-8B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.