Coding ability on OpenAI HumanEval (Score, TPF)
51.22HumanEval ScoreBaseline
Evaluation results
| Method | Links | ||
|---|---|---|---|
| BaselineModel=LLaDA-MoE-Instruct, Gen Length=256, Block Size=64, Early Stop=true2025.10 | 51.22 | 1 | |
| Fast-dLLMModel=LLaDA-MoE-Instruct, Gen Length=256, Block Size=64, Early Stop=true2025.10 | 51.22 | 4.97 | |
| CreditDecodingModel=LLaDA-MoE-Instruct, Gen Length=256, Block Size=64, Early Stop=true2025.10 | 51.22 | 6 | |
| CreditDecodingModel=LLaDA-8B-Instruct, Gen Length=256, Block Size=64, Early Stop=true2025.10 | 36.59 | 4.69 | |
| BaselineModel=LLaDA-8B-Instruct, Gen Length=256, Block Size=64, Early Stop=true2025.10 | 34.76 | 1 | |
| Fast-dLLMModel=LLaDA-8B-Instruct, Gen Length=256, Block Size=64, Early Stop=true2025.10 | 34.15 | 3.82 |