# Mathematical Reasoning on OMNI-MATH (Overall Accuracy) Source: [Mathematical Reasoning on OMNI-MATH (Overall Accuracy)]() Collection index: [llms.txt]() Data updated: 2026-04-15T09:42:46.949Z Evaluation of model performance on the OMNI-MATH dataset for Mathematical Reasoning, reporting the primary overall accuracy 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) | OMNI-MATH | | Metric | Overall Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 25 of 25 | | 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 - [Overall Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Overall Accuracy | Result date | Report paper | | --- | --- | --- | --- | | TEPO | 39.55 | Not recorded | 69df1bbe7d0172140ec7bb01 | | GRPO/DAPO | 38.42 | Not recorded | 69df1bbe7d0172140ec7bb01 | | TEPO | 35.11 | Not recorded | 69df1bbe7d0172140ec7bb01 | | DeepSeek-R1-Distill-Qwen-7B | 33.86 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen3-14B w. GSPO | 33.54 | Not recorded | 69df1bbe7d0172140ec7bb01 | | TEPO | 33.11 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen3-14B w. Entropy-based Term | 33.04 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen3-14B w. GRPO/DAPO | 33.01 | Not recorded | 69df1bbe7d0172140ec7bb01 | | GRPO/DAPO | 32.97 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen3-14B w. CLIP-Cov | 32.94 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen3-14B w. KL-Cov | 32.87 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen3-14B w. GPG | 32.22 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen3-14B | 30.27 | Not recorded | 69df1bbe7d0172140ec7bb01 | | TEPO | 27.17 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen2.5-7B w. GPG | 26.28 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen2.5-7B w. Entropy-based Term | 26.14 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen2.5-7B w. CLIP-Cov | 26.11 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen2.5-7B w. GSPO | 25.72 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen2.5-7B w. KL-Cov | 25.64 | Not recorded | 69df1bbe7d0172140ec7bb01 | | DeepSeek-R1-Distill-Llama-8B | 25.5 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen2.5-7B w. GRPO/DAPO | 25 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Qwen2.5-7B | 13.73 | Not recorded | 69df1bbe7d0172140ec7bb01 | | TEPO | 4.59 | Not recorded | 69df1bbe7d0172140ec7bb01 | | GRPO/DAPO | 4.27 | Not recorded | 69df1bbe7d0172140ec7bb01 | | Mistral-7B-Instruct-v0.2 | 3.98 | Not recorded | 69df1bbe7d0172140ec7bb01 | ### TEPO Overall Accuracy: 39.55. Result ID: 69df5d5a47f34ce924c87f2b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### GRPO/DAPO Overall Accuracy: 38.42. Result ID: 69df5d5a47f34ce924c87f2a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### TEPO Overall Accuracy: 35.11. Result ID: 69df5d5a47f34ce924c87ef9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-14B","Method Variant":"TEPO"}. This record does not assert independent reproduction. ### DeepSeek-R1-Distill-Qwen-7B Overall Accuracy: 33.86. Result ID: 69df5d5a47f34ce924c87f29. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### Qwen3-14B w. GSPO Overall Accuracy: 33.54. Result ID: 69df5d5a47f34ce924c87ef8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-14B","Method Variant":"GSPO"}. This record does not assert independent reproduction. ### TEPO Overall Accuracy: 33.11. Result ID: 69df5d5a47f34ce924c87f25. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Llama-8B"}. This record does not assert independent reproduction. ### Qwen3-14B w. Entropy-based Term Overall Accuracy: 33.04. Result ID: 69df5d5a47f34ce924c87ef6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-14B","Method Variant":"Entropy-based Term"}. This record does not assert independent reproduction. ### Qwen3-14B w. GRPO/DAPO Overall Accuracy: 33.01. Result ID: 69df5d5a47f34ce924c87ef3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen3-14B","Method Variant":"GRPO/DAPO"}. This record does not assert independent reproduction. ### GRPO/DAPO Overall Accuracy: 32.97. Result ID: 69df5d5a47f34ce924c87f24. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Llama-8B"}. This record does not assert independent reproduction. ### Qwen3-14B w. CLIP-Cov Overall Accuracy: 32.94. Result ID: 69df5d5a47f34ce924c87ef4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-14B","Method Variant":"CLIP-Cov"}. This record does not assert independent reproduction. ### Qwen3-14B w. KL-Cov Overall Accuracy: 32.87. Result ID: 69df5d5a47f34ce924c87ef5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-14B","Method Variant":"KL-Cov"}. This record does not assert independent reproduction. ### Qwen3-14B w. GPG Overall Accuracy: 32.22. Result ID: 69df5d5a47f34ce924c87ef7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-14B","Method Variant":"GPG"}. This record does not assert independent reproduction. ### Qwen3-14B Overall Accuracy: 30.27. Result ID: 69df5d5a47f34ce924c87ef2. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-14B","Method Variant":"Base"}. This record does not assert independent reproduction. ### TEPO Overall Accuracy: 27.17. Result ID: 69df5d5a47f34ce924c87ef1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B","Method Variant":"TEPO"}. This record does not assert independent reproduction. ### Qwen2.5-7B w. GPG Overall Accuracy: 26.28. Result ID: 69df5d5a47f34ce924c87eef. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B","Method Variant":"GPG"}. This record does not assert independent reproduction. ### Qwen2.5-7B w. Entropy-based Term Overall Accuracy: 26.14. Result ID: 69df5d5a47f34ce924c87eee. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B","Method Variant":"Entropy-based Term"}. This record does not assert independent reproduction. ### Qwen2.5-7B w. CLIP-Cov Overall Accuracy: 26.11. Result ID: 69df5d5a47f34ce924c87eec. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B","Method Variant":"CLIP-Cov"}. This record does not assert independent reproduction. ### Qwen2.5-7B w. GSPO Overall Accuracy: 25.72. Result ID: 69df5d5a47f34ce924c87ef0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B","Method Variant":"GSPO"}. This record does not assert independent reproduction. ### Qwen2.5-7B w. KL-Cov Overall Accuracy: 25.64. Result ID: 69df5d5a47f34ce924c87eed. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B","Method Variant":"KL-Cov"}. This record does not assert independent reproduction. ### DeepSeek-R1-Distill-Llama-8B Overall Accuracy: 25.5. Result ID: 69df5d5a47f34ce924c87f23. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Llama-8B"}. This record does not assert independent reproduction. ### Qwen2.5-7B w. GRPO/DAPO Overall Accuracy: 25. Result ID: 69df5d5a47f34ce924c87eeb. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B","Method Variant":"GRPO/DAPO"}. This record does not assert independent reproduction. ### Qwen2.5-7B Overall Accuracy: 13.73. Result ID: 69df5d5a47f34ce924c87eea. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B","Method Variant":"Base"}. This record does not assert independent reproduction. ### TEPO Overall Accuracy: 4.59. Result ID: 69df5d5a47f34ce924c87f28. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Mistral-7B-Instruct-v0.2"}. This record does not assert independent reproduction. ### GRPO/DAPO Overall Accuracy: 4.27. Result ID: 69df5d5a47f34ce924c87f27. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Mistral-7B-Instruct-v0.2"}. This record does not assert independent reproduction. ### Mistral-7B-Instruct-v0.2 Overall Accuracy: 3.98. Result ID: 69df5d5a47f34ce924c87f26. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Mistral-7B-Instruct-v0.2"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.