# Mathematical Reasoning on AIME 24 (accuracy, delta) Source: [Mathematical Reasoning on AIME 24 (accuracy, delta)]() Collection index: [llms.txt]() Data updated: 2026-06-11T09:09:18.967Z Evaluation of model performance on the AIME 2024 dataset for mathematical reasoning, reporting Accuracy and the Delta 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) | AIME 24 | | Metric | 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 - [Accuracy](): Higher is better; unit: PERCENT. - [Delta Score](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Berr. Latent | 14 | Not recorded | 69b8e0040636e5bf4afbe967 | | Qwen2.5-Math-7B w/ CADFT | 10.41 | Not recorded | 6a2a40e3d33dfb23e7839135 | | Qwen2.5-Math-1.5B w/ CADFT | 8.76 | Not recorded | 6a2a40e3d33dfb23e7839135 | | Qwen2.5-Math-7B w/ DFT | 8.56 | Not recorded | 6a2a40e3d33dfb23e7839135 | | Qwen2.5-Math-1.5B w/ DFT | 6.87 | Not recorded | 6a2a40e3d33dfb23e7839135 | | Qwen2.5-Math-7B | 6.68 | Not recorded | 6a2a40e3d33dfb23e7839135 | | Qwen2.5-Math-1.5B | 4.16 | Not recorded | 6a2a40e3d33dfb23e7839135 | | LLaDA + Sonnet Plan | 3.3 | Not recorded | 69b8e0040636e5bf4afbe967 | | Qwen2.5-Math-7B w/ SFT | 2.48 | Not recorded | 6a2a40e3d33dfb23e7839135 | | Qwen2.5-Math-1.5B w/ SFT | 1.87 | Not recorded | 6a2a40e3d33dfb23e7839135 | | DeepSeekMath-7B w/ CADFT | 1.65 | Not recorded | 6a2a40e3d33dfb23e7839135 | | Berr. BL | 1.5 | Not recorded | 69b8e0040636e5bf4afbe967 | | Berr. Text | 1.5 | Not recorded | 69b8e0040636e5bf4afbe967 | | DeepSeekMath-7B w/ DFT | 1.24 | Not recorded | 6a2a40e3d33dfb23e7839135 | | LLaMA-3.2-3B w/ CADFT | 1.05 | Not recorded | 6a2a40e3d33dfb23e7839135 | | LLaMA-3.2-3B w/ DFT | 0.83 | Not recorded | 6a2a40e3d33dfb23e7839135 | | LLaMA-3.1-8B w/ CADFT | 0.65 | Not recorded | 6a2a40e3d33dfb23e7839135 | | LLaMA-3.2-3B | 0.41 | Not recorded | 6a2a40e3d33dfb23e7839135 | | LLaMA-3.1-8B w/ DFT | 0.41 | Not recorded | 6a2a40e3d33dfb23e7839135 | | DeepSeekMath-7B w/ SFT | 0.41 | Not recorded | 6a2a40e3d33dfb23e7839135 | | LLaMA-3.1-8B | 0.21 | Not recorded | 6a2a40e3d33dfb23e7839135 | | DeepSeekMath-7B | 0.21 | Not recorded | 6a2a40e3d33dfb23e7839135 | | LLaDA | 0 | Not recorded | 69b8e0040636e5bf4afbe967 | | LLaMA-3.2-3B w/ SFT | 0 | Not recorded | 6a2a40e3d33dfb23e7839135 | | LLaMA-3.1-8B w/ SFT | 0 | Not recorded | 6a2a40e3d33dfb23e7839135 | ### Berr. Latent Accuracy: 14. Result ID: 69b900f547f34ce924a113c5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Collaboration space":"Latent-space","Description":"Best latent-space DDLM→ARM result","Projector training":"trained Linear-GELU-Linear projector on 35K domain-specific samples"}. This record does not assert independent reproduction. ### Qwen2.5-Math-7B w/ CADFT Accuracy: 10.41. Result ID: 6a2a7b3bfbc373592c27d5a8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B","Fine-tuning Protocol":"CADFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### Qwen2.5-Math-1.5B w/ CADFT Accuracy: 8.76. Result ID: 6a2a7b3bfbc373592c27d5a4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-1.5B","Fine-tuning Protocol":"CADFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### Qwen2.5-Math-7B w/ DFT Accuracy: 8.56. Result ID: 6a2a7b3bfbc373592c27d5a7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B","Fine-tuning Protocol":"DFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### Qwen2.5-Math-1.5B w/ DFT Accuracy: 6.87. Result ID: 6a2a7b3bfbc373592c27d5a3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-1.5B","Fine-tuning Protocol":"DFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### Qwen2.5-Math-7B Accuracy: 6.68. Result ID: 6a2a7b3bfbc373592c27d5a5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B","Fine-tuning Protocol":"None","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### Qwen2.5-Math-1.5B Accuracy: 4.16. Result ID: 6a2a7b3bfbc373592c27d5a1. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-1.5B","Fine-tuning Protocol":"None","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### LLaDA + Sonnet Plan Accuracy: 3.3. Result ID: 69b900f547f34ce924a113c2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Plan conditioning":"Sonnet Plan"}. This record does not assert independent reproduction. ### Qwen2.5-Math-7B w/ SFT Accuracy: 2.48. Result ID: 6a2a7b3bfbc373592c27d5a6. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-7B","Fine-tuning Protocol":"SFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### Qwen2.5-Math-1.5B w/ SFT Accuracy: 1.87. Result ID: 6a2a7b3bfbc373592c27d5a2. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-Math-1.5B","Fine-tuning Protocol":"SFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### DeepSeekMath-7B w/ CADFT Accuracy: 1.65. Result ID: 6a2a7b3bfbc373592c27d5a0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeekMath-7B","Fine-tuning Protocol":"CADFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### Berr. BL Accuracy: 1.5. Result ID: 69b900f547f34ce924a113c3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Description":"LLaDA-only baseline by Berrayana et al."}. This record does not assert independent reproduction. ### Berr. Text Accuracy: 1.5. Result ID: 69b900f547f34ce924a113c4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Collaboration space":"Text-space","Description":"Best text-space collaboration result"}. This record does not assert independent reproduction. ### DeepSeekMath-7B w/ DFT Accuracy: 1.24. Result ID: 6a2a7b3bfbc373592c27d59f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DeepSeekMath-7B","Fine-tuning Protocol":"DFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### LLaMA-3.2-3B w/ CADFT Accuracy: 1.05. Result ID: 6a2a7b3bfbc373592c27d598. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaMA-3.2-3B","Fine-tuning Protocol":"CADFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### LLaMA-3.2-3B w/ DFT Accuracy: 0.83. Result ID: 6a2a7b3bfbc373592c27d597. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaMA-3.2-3B","Fine-tuning Protocol":"DFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### LLaMA-3.1-8B w/ CADFT Accuracy: 0.65. Result ID: 6a2a7b3bfbc373592c27d59c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B","Fine-tuning Protocol":"CADFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### LLaMA-3.2-3B Accuracy: 0.41. Result ID: 6a2a7b3bfbc373592c27d595. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaMA-3.2-3B","Fine-tuning Protocol":"None","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### LLaMA-3.1-8B w/ DFT Accuracy: 0.41. Result ID: 6a2a7b3bfbc373592c27d59b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B","Fine-tuning Protocol":"DFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### DeepSeekMath-7B w/ SFT Accuracy: 0.41. Result ID: 6a2a7b3bfbc373592c27d59e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"DeepSeekMath-7B","Fine-tuning Protocol":"SFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### LLaMA-3.1-8B Accuracy: 0.21. Result ID: 6a2a7b3bfbc373592c27d599. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B","Fine-tuning Protocol":"None","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### DeepSeekMath-7B Accuracy: 0.21. Result ID: 6a2a7b3bfbc373592c27d59d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"DeepSeekMath-7B","Fine-tuning Protocol":"None","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### LLaDA Accuracy: 0. Result ID: 69b900f547f34ce924a113c1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Description":"LLaDA-only baseline"}. This record does not assert independent reproduction. ### LLaMA-3.2-3B w/ SFT Accuracy: 0. Result ID: 6a2a7b3bfbc373592c27d596. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaMA-3.2-3B","Fine-tuning Protocol":"SFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. ### LLaMA-3.1-8B w/ SFT Accuracy: 0. Result ID: 6a2a7b3bfbc373592c27d59a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaMA-3.1-8B","Fine-tuning Protocol":"SFT","Evaluation Protocol":"Average@16"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.