List your company

Code Generation on LiveCodeBench (out-of-distribution)

66AccuracyHINT TUNING
36.15243.90151.6559.399May 9, 2026

Evaluation results

MethodLinks
HINT TUNINGBackbone=DS-R1-32B2026.056613
SFT-S1.1Backbone=DS-R1-32B2026.0565.711.8
OriginalBackbone=Qwen3-4B2026.0562.717.4
HINT TUNINGBackbone=Qwen3-4B2026.0561.913.2
SFT-S1.1Backbone=DS-R1-14B2026.0557.112.1
OriginalBackbone=DS-R1-32B2026.0556.711.1
HINT TUNINGBackbone=DS-R1-14B2026.0553.713.8
SFT-S1.1Backbone=Qwen3-4B2026.0550.414.7
OriginalBackbone=DS-R1-14B2026.0549.310.8
SFT-S1.1Backbone=DS-R1-7B2026.0539.215
HINT TUNINGBackbone=DS-R1-7B2026.0537.716.6
OriginalBackbone=DS-R1-7B2026.0537.311.8