# Software Engineering Question Answering on SWE-QA Pro Source: [Software Engineering Question Answering on SWE-QA Pro]() Collection index: [llms.txt]() Data updated: 2026-05-11T08:03:43.320Z Evaluation of model performance on the SWE-QA Pro dataset for Software Engineering Question Answering tasks, measured using Rubric Judge 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 | Software Engineering Question Answering | | Dataset (including any stated split) | SWE-QA Pro | | Metric | Rubric Judge Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 15 of 15 | | 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 - [Rubric Judge Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Rubric Judge Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Predicate-Based Belief State | 51.4 | Not recorded | 6a016295dbc836b87d823143 | | Baseline | 50.9 | Not recorded | 6a016295dbc836b87d823143 | | Baseline | 48.9 | Not recorded | 6a016295dbc836b87d823143 | | Predicate-Based Belief State | 48.4 | Not recorded | 6a016295dbc836b87d823143 | | Predicate-Based Belief State | 47.7 | Not recorded | 6a016295dbc836b87d823143 | | Lobotomized | 43.6 | Not recorded | 6a016295dbc836b87d823143 | | Predicate-Based Belief State | 43 | Not recorded | 6a016295dbc836b87d823143 | | Predicate-Based Belief State | 42.3 | Not recorded | 6a016295dbc836b87d823143 | | Baseline | 39.8 | Not recorded | 6a016295dbc836b87d823143 | | Predicate-Based Belief State | 37.3 | Not recorded | 6a016295dbc836b87d823143 | | Predicate-Based Belief State | 34.6 | Not recorded | 6a016295dbc836b87d823143 | | Lobotomized | 33.7 | Not recorded | 6a016295dbc836b87d823143 | | Predicate-Based Belief State | 33.3 | Not recorded | 6a016295dbc836b87d823143 | | Lobotomized | 31.7 | Not recorded | 6a016295dbc836b87d823143 | | Baseline | 27.3 | Not recorded | 6a016295dbc836b87d823143 | ### Predicate-Based Belief State Rubric Judge Accuracy: 51.4. Result ID: 6a018d4a47f34ce924e3345f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Harness":"Iter-RetGen","Representation":"Freeform (b\_free)"}. This record does not assert independent reproduction. ### Baseline Rubric Judge Accuracy: 50.9. Result ID: 6a018d4a47f34ce924e3345d. [Reporting paper]() Recorded evaluation configuration: {"Harness":"Iter-RetGen"}. This record does not assert independent reproduction. ### Baseline Rubric Judge Accuracy: 48.9. Result ID: 6a018d4a47f34ce924e33460. [Reporting paper]() Recorded evaluation configuration: {"Harness":"MemGPT"}. This record does not assert independent reproduction. ### Predicate-Based Belief State Rubric Judge Accuracy: 48.4. Result ID: 6a018d4a47f34ce924e3345e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Harness":"Iter-RetGen","Representation":"Structured JSON (b\_struct)"}. This record does not assert independent reproduction. ### Predicate-Based Belief State Rubric Judge Accuracy: 47.7. Result ID: 6a018d4a47f34ce924e33463. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Harness":"MemGPT","Representation":"Freeform (b\_free)"}. This record does not assert independent reproduction. ### Lobotomized Rubric Judge Accuracy: 43.6. Result ID: 6a018d4a47f34ce924e33461. [Reporting paper]() Recorded evaluation configuration: {"Harness":"MemGPT"}. This record does not assert independent reproduction. ### Predicate-Based Belief State Rubric Judge Accuracy: 43. Result ID: 6a018d4a47f34ce924e3345c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Harness":"ReAct","Representation":"Freeform (b\_free)"}. This record does not assert independent reproduction. ### Predicate-Based Belief State Rubric Judge Accuracy: 42.3. Result ID: 6a018d4a47f34ce924e33462. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Harness":"MemGPT","Representation":"Structured JSON (b\_struct)"}. This record does not assert independent reproduction. ### Baseline Rubric Judge Accuracy: 39.8. Result ID: 6a018d4a47f34ce924e33459. [Reporting paper]() Recorded evaluation configuration: {"Harness":"ReAct"}. This record does not assert independent reproduction. ### Predicate-Based Belief State Rubric Judge Accuracy: 37.3. Result ID: 6a018d4a47f34ce924e3345b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Harness":"ReAct","Representation":"Structured JSON (b\_struct)"}. This record does not assert independent reproduction. ### Predicate-Based Belief State Rubric Judge Accuracy: 34.6. Result ID: 6a018d4a47f34ce924e33458. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Harness":"IRCoT","Representation":"Freeform (b\_free)"}. This record does not assert independent reproduction. ### Lobotomized Rubric Judge Accuracy: 33.7. Result ID: 6a018d4a47f34ce924e3345a. [Reporting paper]() Recorded evaluation configuration: {"Harness":"ReAct"}. This record does not assert independent reproduction. ### Predicate-Based Belief State Rubric Judge Accuracy: 33.3. Result ID: 6a018d4a47f34ce924e33457. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Harness":"IRCoT","Representation":"Structured JSON (b\_struct)"}. This record does not assert independent reproduction. ### Lobotomized Rubric Judge Accuracy: 31.7. Result ID: 6a018d4a47f34ce924e33456. [Reporting paper]() Recorded evaluation configuration: {"Harness":"IRCoT"}. This record does not assert independent reproduction. ### Baseline Rubric Judge Accuracy: 27.3. Result ID: 6a018d4a47f34ce924e33455. [Reporting paper]() Recorded evaluation configuration: {"Harness":"IRCoT"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.