# Mathematical Reasoning on OlympiadBench (test) (Acc, Len, AE) Source: [Mathematical Reasoning on OlympiadBench (test) (Acc, Len, AE)]() Collection index: [llms.txt]() Data updated: 2026-05-22T08:31:03.772Z Evaluation of model performance on mathematical reasoning problems using the test split of OlympiadBench, measured by Accuracy (Acc.), output Length (Len.), and Answer Error (AE). 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) | OlympiadBench (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 18 of 18 | | 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. - [Output Length](): Higher is better; unit: SCALAR. - [Answer Error](): Lower is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | GRPO+CLORE | 43.2 | Not recorded | 6a0fe2fdf984739ed6739c92 | | DAPO | 42.9 | Not recorded | 6a0fe2fdf984739ed6739c92 | | DAPO+CLORE | 42.9 | Not recorded | 6a0fe2fdf984739ed6739c92 | | Training Efficient | 42.8 | Not recorded | 6a0fe2fdf984739ed6739c92 | | ThinkPrune+CLORE | 42.8 | Not recorded | 6a0fe2fdf984739ed6739c92 | | GRPO | 42.1 | Not recorded | 6a0fe2fdf984739ed6739c92 | | Training Efficient+CLORE | 41.5 | Not recorded | 6a0fe2fdf984739ed6739c92 | | ThinkPrune | 39.7 | Not recorded | 6a0fe2fdf984739ed6739c92 | | GRPO+CLORE | 38.3 | Not recorded | 6a0fe2fdf984739ed6739c92 | | Base model (w/o training) | 38 | Not recorded | 6a0fe2fdf984739ed6739c92 | | ThinkPrune | 36.4 | Not recorded | 6a0fe2fdf984739ed6739c92 | | DAPO+CLORE | 35.7 | Not recorded | 6a0fe2fdf984739ed6739c92 | | Training Efficient+CLORE | 35.3 | Not recorded | 6a0fe2fdf984739ed6739c92 | | ThinkPrune+CLORE | 34.2 | Not recorded | 6a0fe2fdf984739ed6739c92 | | GRPO | 32.5 | Not recorded | 6a0fe2fdf984739ed6739c92 | | Training Efficient | 32.5 | Not recorded | 6a0fe2fdf984739ed6739c92 | | DAPO | 28.5 | Not recorded | 6a0fe2fdf984739ed6739c92 | | Base model (w/o training) | 16.1 | Not recorded | 6a0fe2fdf984739ed6739c92 | ### GRPO+CLORE Accuracy: 43.2. Result ID: 6a101441fbc373592c090fae. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### DAPO Accuracy: 42.9. Result ID: 6a101441fbc373592c090faf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### DAPO+CLORE Accuracy: 42.9. Result ID: 6a101441fbc373592c090fb0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### Training Efficient Accuracy: 42.8. Result ID: 6a101441fbc373592c090fb1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### ThinkPrune+CLORE Accuracy: 42.8. Result ID: 6a101441fbc373592c090fb4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### GRPO Accuracy: 42.1. Result ID: 6a101441fbc373592c090fad. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### Training Efficient+CLORE Accuracy: 41.5. Result ID: 6a101441fbc373592c090fb2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### ThinkPrune Accuracy: 39.7. Result ID: 6a101441fbc373592c090fb3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### GRPO+CLORE Accuracy: 38.3. Result ID: 6a101441fbc373592c090fb7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction. ### Base model (w/o training) Accuracy: 38. Result ID: 6a101441fbc373592c090fac. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"DeepSeek-R1-Distill-Qwen-7B"}. This record does not assert independent reproduction. ### ThinkPrune Accuracy: 36.4. Result ID: 6a101441fbc373592c090fbc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction. ### DAPO+CLORE Accuracy: 35.7. Result ID: 6a101441fbc373592c090fb9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction. ### Training Efficient+CLORE Accuracy: 35.3. Result ID: 6a101441fbc373592c090fbb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction. ### ThinkPrune+CLORE Accuracy: 34.2. Result ID: 6a101441fbc373592c090fbd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction. ### GRPO Accuracy: 32.5. Result ID: 6a101441fbc373592c090fb6. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction. ### Training Efficient Accuracy: 32.5. Result ID: 6a101441fbc373592c090fba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction. ### DAPO Accuracy: 28.5. Result ID: 6a101441fbc373592c090fb8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction. ### Base model (w/o training) Accuracy: 16.1. Result ID: 6a101441fbc373592c090fb5. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.