| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| WebNLG+ 2020 3.0 (testB) | ReGen | BLEU0.563 | 10 | 4d ago | |
| WebNLG OOD standard (test) | Rule-based NLG (trained by Qwen 3 235B) | BLEU37.72 | 6 | 4d ago | |
| WebNLG All standard (test) | Fine-tuned BART | BLEU0.4352 | 6 | 4d ago | |
| OpenDialKG (test) | Llama 3.3 | Grammaticality98.5 | 6 | 4d ago | |
| GEM2 Factual (test) | Llama 3.3 | Grammaticality98.4 | 6 | 4d ago | |
| GEM2 Fictional (test) | Llama 3.3 | Grammaticality93.7 | 6 | 4d ago | |
| GEM2 Counterfactual (test) | Llama 3.3 | Grammaticality81.8 | 6 | 4d ago | |
| WebNLG (test) | Llama 3.3 | Grammaticality75.2 | 6 | 4d ago | |
| WebNLG Unconstrained 2.0 (test) | JointGT (T5) | BLEU0.6614 | 6 | 4d ago | |
| WebNLG Constrained 2.0 (test) | JointGT (T5) | BLEU61.01 | 5 | 4d ago |