Task-oriented Plan Retrieval on Average across domains Single-turn
83Pass RateSemantic XPath
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Semantic XPathModel=GPT-5 mini, Scorer=Entail2026.03 | 83 | 5,691 | |
| In-contextModel=Gemini 3 Flash, Scorer=–2026.03 | 79 | 25,387 | |
| Semantic XPathModel=Gemini 3 Flash, Scorer=Entail2026.03 | 70.7 | 5,830 | |
| In-contextModel=GPT-5 mini, Scorer=–2026.03 | 70 | 26,216 | |
| Semantic XPathModel=GPT-5 mini, Scorer=Sem2026.03 | 66.3 | 6,166 | |
| Semantic XPathModel=Gemini 3 Flash, Scorer=Sem2026.03 | 64.3 | 6,043 | |
| Flat RAGModel=Gemini 3 Flash, Scorer=–2026.03 | 37.7 | 3,361 | |
| Flat RAGModel=GPT-5 mini, Scorer=–2026.03 | 30 | 3,250 |