Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Open-ended text generation on Arena-hard Hard-Prompt
Loading...
58.5
Pairwise Win Rate
LLaDA-1.5
25.948
34.399
42.85
51.301
Jan 30, 2026
Pairwise Win Rate
Updated 4d ago
Evaluation Results
Method
Method
Links
Pairwise Win Rate
LLaDA-1.5
Base Model=LLaDA-1.5
2026.01
58.5
CTC-trained MDLM
Training Objective=CTC...
2026.01
51.4
LLaDA-8B-Instruct
Base Model=LLaDA-8B-In...
2026.01
50
CE-only baseline
Training Objective=CE...
2026.01
27.2
Feedback
Search any
task
Search any
task