Reasoning Performance (Aggregate) on AVG
351TPFLLaDA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LLaDAModel Type=Masked Diffusion2026.02 | 351 | 13.7 | |
| LLaDA-1.5Model Type=Masked Diffusion2026.02 | 348 | 13.5 | |
| TDAR-8B-thinking + BACDModel Type=Block Diffusion, Block Size (B)=16, Decoding=BACD2026.02 | 337 | 51.2 | |
| TDAR-8B-thinkingModel Type=Block Diffusion, Block Size (B)=162026.02 | 297 | 50.5 | |
| LLaDA-MoEModel Type=Masked Diffusion2026.02 | 262 | 22.2 | |
| Fast-dLLM-v2Model Type=Block Diffusion, Block Size (B)=322026.02 | 243 | 19.9 | |
| SDAR-8B-ChatModel Type=Block Diffusion, Block Size (B)=42026.02 | 221 | 17.6 | |
| DiRL-8B-InstructModel Type=Block Diffusion, Block Size (B)=42026.02 | 219 | 38.9 | |
| TDAR-8B-thinking + BACD + TCCFModel Type=Block Diffusion, Block Size (B)=16, Decoding=BACD + TCCF2026.02 | 216 | 55.9 | |
| TraDo-8B-InstructModel Type=Block Diffusion, Block Size (B)=42026.02 | 193 | 31.8 | |
| TraDo-8B-Thinking + BACDModel Type=Block Diffusion, Block Size (B)=4, Decoding=BACD2026.02 | 133 | 48.7 | |
| TraDo-8B-ThinkingModel Type=Block Diffusion, Block Size (B)=42026.02 | 127 | 47.1 | |
| TraDo-8B-Thinking + BACD + TCCFModel Type=Block Diffusion, Block Size (B)=4, Decoding=BACD + TCCF2026.02 | 127 | 49 | |
| Qwen3-8B-Thinking†Model Type=Autoregressive2026.02 | 100 | 67.4 |