# Short-Form Question Answering on SAGE Web Search Source: [Short-Form Question Answering on SAGE Web Search]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:34:13.804Z Evaluation of model performance on the SAGE Web Search dataset, focusing on Short-Form Question Answering, with results broken down by subject domain (e.g., Computer Science, Healthcare). 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 | Short-Form Question Answering | | Dataset (including any stated split) | SAGE Web Search | | Metric | Accuracy (Com. Sci.) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 12 of 12 | | 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 (Com. Sci.)](): Higher is better; unit: PERCENT. - [Accuracy (Healthcare)](): Higher is better; unit: PERCENT. - [Accuracy (Humanities)](): Higher is better; unit: PERCENT. - [Accuracy (Nat. Sci.)](): Higher is better; unit: PERCENT. - [Average Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy (Com. Sci.) | Result date | Report paper | | --- | --- | --- | --- | | BM25 | 63.3 | Not recorded | 698575dfc97a0b3e3b5d5eec | | GPT-5 | 57.3 | Not recorded | 698575dfc97a0b3e3b5d5eec | | BM25 | 56 | Not recorded | 698575dfc97a0b3e3b5d5eec | | gte-Qwen | 44.4 | Not recorded | 698575dfc97a0b3e3b5d5eec | | GPT-5-mini | 40 | Not recorded | 698575dfc97a0b3e3b5d5eec | | DR Tulu | 36 | Not recorded | 698575dfc97a0b3e3b5d5eec | | gte-Qwen | 32.9 | Not recorded | 698575dfc97a0b3e3b5d5eec | | GPT-5-nano | 30.7 | Not recorded | 698575dfc97a0b3e3b5d5eec | | Gemini-2.5-flash | 30.3 | Not recorded | 698575dfc97a0b3e3b5d5eec | | ReasonIR | 28 | Not recorded | 698575dfc97a0b3e3b5d5eec | | Gemini-2.5-pro | 27.7 | Not recorded | 698575dfc97a0b3e3b5d5eec | | ReasonIR | 25.9 | Not recorded | 698575dfc97a0b3e3b5d5eec | ### BM25 Accuracy (Com. Sci.): 63.3. Result ID: 6998a38c5698665ec406121f. [Reporting paper]() Recorded evaluation configuration: {"k":"10","Avg. Searches":"6.42","Avg. Refs":"40.0"}. This record does not assert independent reproduction. ### GPT-5 Accuracy (Com. Sci.): 57.3. Result ID: 6998a38c5698665ec4061213. [Reporting paper]() Recorded evaluation configuration: {"Avg. Searches":"8.78","Avg. Refs":"6.54"}. This record does not assert independent reproduction. ### BM25 Accuracy (Com. Sci.): 56. Result ID: 6998a38c5698665ec4061220. [Reporting paper]() Recorded evaluation configuration: {"k":"5","Avg. Searches":"7.54","Avg. Refs":"22.4"}. This record does not assert independent reproduction. ### gte-Qwen Accuracy (Com. Sci.): 44.4. Result ID: 6998a38c5698665ec4061221. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"10","Avg. Searches":"5.88","Avg. Refs":"33.2"}. This record does not assert independent reproduction. ### GPT-5-mini Accuracy (Com. Sci.): 40. Result ID: 6998a38c5698665ec4061214. [Reporting paper]() Recorded evaluation configuration: {"Avg. Searches":"8.15","Avg. Refs":"4.69"}. This record does not assert independent reproduction. ### DR Tulu Accuracy (Com. Sci.): 36. Result ID: 6998a38c5698665ec4061216. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Avg. Searches":"7.35","Avg. Refs":"37.32"}. This record does not assert independent reproduction. ### gte-Qwen Accuracy (Com. Sci.): 32.9. Result ID: 6998a38c5698665ec4061223. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"5","Avg. Searches":"4.82","Avg. Refs":"14.2"}. This record does not assert independent reproduction. ### GPT-5-nano Accuracy (Com. Sci.): 30.7. Result ID: 6998a38c5698665ec4061215. [Reporting paper]() Recorded evaluation configuration: {"Avg. Searches":"8.92","Avg. Refs":"7.06"}. This record does not assert independent reproduction. ### Gemini-2.5-flash Accuracy (Com. Sci.): 30.3. Result ID: 6998a38c5698665ec4061218. [Reporting paper]() Recorded evaluation configuration: {"Avg. Searches":"15.64","Avg. Refs":"3.79"}. This record does not assert independent reproduction. ### ReasonIR Accuracy (Com. Sci.): 28. Result ID: 6998a38c5698665ec4061222. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"10","Avg. Searches":"7.51","Avg. Refs":"35.8"}. This record does not assert independent reproduction. ### Gemini-2.5-pro Accuracy (Com. Sci.): 27.7. Result ID: 6998a38c5698665ec4061217. [Reporting paper]() Recorded evaluation configuration: {"Avg. Searches":"14.02","Avg. Refs":"1.43"}. This record does not assert independent reproduction. ### ReasonIR Accuracy (Com. Sci.): 25.9. Result ID: 6998a38c5698665ec4061224. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"k":"5","Avg. Searches":"8.78","Avg. Refs":"17.4"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.