Reasoning on SQuAD 2.0
90.8ScoreLLaDA2.1-flash
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LLaDA2.1-flashInference Mode=Q Mode2026.02 | 90.8 | 390 | |
| LLaDA2.1-flashInference Mode=S Mode2026.02 | 90.65 | 501 | |
| LLaDA2.0-flash2026.02 | 90 | 310 | |
| Qwen3-30B-A3B-Inst-25072026.02 | 89.51 | 100 | |
| LLaDA2.0-mini2026.02 | 86.5 | 2.47 | |
| Qwen3-8Bno think=true2026.02 | 85.21 | — | |
| LLaDA2.1-minimode=Q Mode2026.02 | 85.13 | 3.09 | |
| LLaDA2.1-minimode=S Mode2026.02 | 84.51 | 4.33 | |
| Ling-flash-2.02026.02 | 81.32 | 100 | |
| Ling-mini-2.02026.02 | 75.56 | — |