List your company

Logical Reasoning on Sudoku (Token Length Specific)

23.88Accuracy (128 tokens)AIS
21.5422.147522.75523.3625May 13, 2026

Evaluation results

MethodLinks
AISBackbone=LLaDA-8B, Evaluation Protocol=Full-model fine-tuning2026.0523.8825.0724.07
FlashRL (TIS)Backbone=LLaDA-8B, Evaluation Protocol=Full-model fine-tuning2026.0523.824.6923.73
BF16Backbone=LLaDA-8B, Evaluation Protocol=Full-model fine-tuning2026.0522.126.729.25
FP8 RolloutBackbone=LLaDA-8B, Evaluation Protocol=Full-model fine-tuning2026.0521.6323.323.01