# Mathematical Reasoning on AIME 2026 (Accuracy, Average tokens generated) Source: [Mathematical Reasoning on AIME 2026 (Accuracy, Average tokens generated)]() Collection index: [llms.txt]() Data updated: 2026-06-11T06:14:44.378Z Evaluation of model performance on the American Invitational Mathematics Examination (AIME) 2026 dataset, reporting Accuracy and the average number of tokens generated per response. 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) | AIME 2026 | | 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. - [Average Tokens Generated](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Gemma-4 31B | 86.67 | Not recorded | 6a28f33c90966ade919c16f3 | | Gemma-4 31B | 83.33 | Not recorded | 6a28f33c90966ade919c16f3 | | Gemma-4 31B | 83.33 | Not recorded | 6a28f33c90966ade919c16f3 | | Phi-4 14B | 80 | Not recorded | 6a28f33c90966ade919c16f3 | | Gemma-4 31B | 80 | Not recorded | 6a28f33c90966ade919c16f3 | | Phi-4 14B | 76.67 | Not recorded | 6a28f33c90966ade919c16f3 | | Qwen3 4B | 73.33 | Not recorded | 6a28f33c90966ade919c16f3 | | Qwen3 4B | 70 | Not recorded | 6a28f33c90966ade919c16f3 | | Qwen3 4B | 53.33 | Not recorded | 6a28f33c90966ade919c16f3 | | Gemma-4 E4B | 53.33 | Not recorded | 6a28f33c90966ade919c16f3 | | Gemma-4 E4B | 46.67 | Not recorded | 6a28f33c90966ade919c16f3 | | Gemma-4 E4B | 43.33 | Not recorded | 6a28f33c90966ade919c16f3 | | Qwen3 4B | 40 | Not recorded | 6a28f33c90966ade919c16f3 | | Gemma-4 E4B | 33.33 | Not recorded | 6a28f33c90966ade919c16f3 | | Phi-4 14B | 16.67 | Not recorded | 6a28f33c90966ade919c16f3 | | Phi-4 14B | 10 | Not recorded | 6a28f33c90966ade919c16f3 | ### Gemma-4 31B Accuracy: 86.67. Result ID: 6a2a518afbc373592c269a3d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Reasoning+Rotate2Think"}. This record does not assert independent reproduction. ### Gemma-4 31B Accuracy: 83.33. Result ID: 6a2a518afbc373592c269a3b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Base+Rotate2Think"}. This record does not assert independent reproduction. ### Gemma-4 31B Accuracy: 83.33. Result ID: 6a2a518afbc373592c269a3c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Reasoning"}. This record does not assert independent reproduction. ### Phi-4 14B Accuracy: 80. Result ID: 6a2a518afbc373592c269a39. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Reasoning+Rotate2Think"}. This record does not assert independent reproduction. ### Gemma-4 31B Accuracy: 80. Result ID: 6a2a518afbc373592c269a3a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Base"}. This record does not assert independent reproduction. ### Phi-4 14B Accuracy: 76.67. Result ID: 6a2a518afbc373592c269a38. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Reasoning"}. This record does not assert independent reproduction. ### Qwen3 4B Accuracy: 73.33. Result ID: 6a2a518afbc373592c269a31. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Reasoning+Rotate2Think"}. This record does not assert independent reproduction. ### Qwen3 4B Accuracy: 70. Result ID: 6a2a518afbc373592c269a30. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Reasoning"}. This record does not assert independent reproduction. ### Qwen3 4B Accuracy: 53.33. Result ID: 6a2a518afbc373592c269a2f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Base+Rotate2Think"}. This record does not assert independent reproduction. ### Gemma-4 E4B Accuracy: 53.33. Result ID: 6a2a518afbc373592c269a35. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Reasoning+Rotate2Think"}. This record does not assert independent reproduction. ### Gemma-4 E4B Accuracy: 46.67. Result ID: 6a2a518afbc373592c269a33. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Base+Rotate2Think"}. This record does not assert independent reproduction. ### Gemma-4 E4B Accuracy: 43.33. Result ID: 6a2a518afbc373592c269a34. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Reasoning"}. This record does not assert independent reproduction. ### Qwen3 4B Accuracy: 40. Result ID: 6a2a518afbc373592c269a2e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Base"}. This record does not assert independent reproduction. ### Gemma-4 E4B Accuracy: 33.33. Result ID: 6a2a518afbc373592c269a32. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Base"}. This record does not assert independent reproduction. ### Phi-4 14B Accuracy: 16.67. Result ID: 6a2a518afbc373592c269a37. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Base+Rotate2Think"}. This record does not assert independent reproduction. ### Phi-4 14B Accuracy: 10. Result ID: 6a2a518afbc373592c269a36. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Mode":"Base"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.