# Factual Question Answering on SimpleQA (test) Source: [Factual Question Answering on SimpleQA (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:32:48.222Z Evaluation of model performance on the test split of the SimpleQA dataset for Factual Question Answering tasks, measured using Accuracy. 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 | Factual Question Answering | | Dataset (including any stated split) | SimpleQA (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 10 of 10 | | 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. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | HEART | 79.07 | Not recorded | 68fa7be71e209131cea8b345 | | Wait | 73.61 | Not recorded | 68fa7be71e209131cea8b345 | | Self Reflection | 72.91 | Not recorded | 68fa7be71e209131cea8b345 | | CoT | 72.73 | Not recorded | 68fa7be71e209131cea8b345 | | Vanilla | 70.7 | Not recorded | 68fa7be71e209131cea8b345 | | HEART | 59.85 | Not recorded | 68fa7be71e209131cea8b345 | | CoT | 58.38 | Not recorded | 68fa7be71e209131cea8b345 | | Wait | 55.1 | Not recorded | 68fa7be71e209131cea8b345 | | Self Reflection | 53.1 | Not recorded | 68fa7be71e209131cea8b345 | | Vanilla | 34.19 | Not recorded | 68fa7be71e209131cea8b345 | ### HEART Accuracy: 79.07. Result ID: 699a53f85698665ec4db9f27. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Gemini 3 Pro","Prompt Strategy":"HEART","Setting":"Human-in-the-Loop Proxy"}. This record does not assert independent reproduction. ### Wait Accuracy: 73.61. Result ID: 699a53f85698665ec4db9f26. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"Gemini 3 Pro","Prompt Strategy":"Wait","Setting":"Human-in-the-Loop Proxy"}. This record does not assert independent reproduction. ### Self Reflection Accuracy: 72.91. Result ID: 699a53f85698665ec4db9f24. [Reporting paper]() Recorded evaluation configuration: {"Model":"Gemini 3 Pro","Prompt Strategy":"Self Reflection","Setting":"Human-in-the-Loop Proxy"}. This record does not assert independent reproduction. ### CoT Accuracy: 72.73. Result ID: 699a53f85698665ec4db9f25. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Gemini 3 Pro","Prompt Strategy":"CoT","Setting":"Human-in-the-Loop Proxy"}. This record does not assert independent reproduction. ### Vanilla Accuracy: 70.7. Result ID: 699a53f85698665ec4db9f23. [Reporting paper]() Recorded evaluation configuration: {"Model":"Gemini 3 Pro","Prompt Strategy":"Vanilla","Setting":"Human-in-the-Loop Proxy"}. This record does not assert independent reproduction. ### HEART Accuracy: 59.85. Result ID: 699a53f85698665ec4db9f22. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Gemini 3 Flash","Prompt Strategy":"HEART","Setting":"Human-in-the-Loop Proxy"}. This record does not assert independent reproduction. ### CoT Accuracy: 58.38. Result ID: 699a53f85698665ec4db9f20. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Gemini 3 Flash","Prompt Strategy":"CoT","Setting":"Human-in-the-Loop Proxy"}. This record does not assert independent reproduction. ### Wait Accuracy: 55.1. Result ID: 699a53f85698665ec4db9f21. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"Gemini 3 Flash","Prompt Strategy":"Wait","Setting":"Human-in-the-Loop Proxy"}. This record does not assert independent reproduction. ### Self Reflection Accuracy: 53.1. Result ID: 699a53f85698665ec4db9f1f. [Reporting paper]() Recorded evaluation configuration: {"Model":"Gemini 3 Flash","Prompt Strategy":"Self Reflection","Setting":"Human-in-the-Loop Proxy"}. This record does not assert independent reproduction. ### Vanilla Accuracy: 34.19. Result ID: 699a53f85698665ec4db9f1e. [Reporting paper]() Recorded evaluation configuration: {"Model":"Gemini 3 Flash","Prompt Strategy":"Vanilla","Setting":"Human-in-the-Loop Proxy"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.