Knowledge & Instruction Following on ARC Challenge
97.7Accuracy (ARC)Qwen3.5-9B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3.5-9BParams=9B, Sampling mode=AR2026.06 | 97.7 | |
| FLARE-9BParams=9B, Sampling mode=AR-Trust2026.06 | 96.33 | |
| Qwen3.5Params=4B, Sampling mode=AR2026.06 | 96.33 | |
| LLaDA-2.0-flashParams=100B-A5B, Sampling mode=Diffusion2026.06 | 95.93 | |
| FLARE-9BParams=9B, Sampling mode=Diffusion-Trust2026.06 | 95.65 | |
| FLAREParams=4B, Sampling mode=Diffusion-Trust2026.06 | 94.62 | |
| LLaDA-2.0-miniParams=16B-A1B, Sampling mode=Diffusion2026.06 | 93.56 | |
| FLAREParams=4B, Sampling mode=AR-Trust2026.06 | 93.52 | |
| SDAR 30B-A3BParams=30B-A3B, Sampling mode=Diffusion2026.06 | 93.2 | |
| Qwen3.5Params=2B, Sampling mode=AR2026.06 | 92.15 | |
| SDARParams=8B, Sampling mode=Diffusion2026.06 | 91.9 | |
| SDARParams=4B, Sampling mode=Diffusion2026.06 | 90.5 | |
| FLAREParams=2B, Sampling mode=Diffusion-Trust2026.06 | 85.84 | |
| SDARParams=1.7B, Sampling mode=Diffusion2026.06 | 85.4 | |
| FLAREParams=2B, Sampling mode=AR-Trust2026.06 | 85.07 |