Turning the TIDE: Cross-Architecture Distillation for Diffusion Large Language Models
Open paper
Benchmark
Model
Rank
Results
code-generation-on-humaneval
Qwen3-0.6B-BD3LM (Tide-Shared)
#105
Pass@1: 49.39
code-generation-on-mbpp
Qwen3-0.6B-BD3LM (Tide-Shared)
#138
Accuracy: 38.6
Rank counts only results with a code link.