List your company

Code Reasoning on LiveCodeBench (Pass@1)

88.3Pass@1 AccuracyREBALANCE
16.74835.32453.972.476Nov 5, 2025Nov 26, 2025Dec 17, 2025Jan 7, 2026Jan 28, 2026Feb 18, 2026Mar 12, 2026

Evaluation results

MethodLinks
REBALANCEBackbone=QwQ-32B2026.0388.35,000
BaselineBackbone=QwQ-32B2026.0387.56,000
REBALANCEBackbone=Qwen3-14B2026.0384.66,000
BaselineBackbone=Qwen3-14B2026.0383.57,000
SnapStreamMaximum sequence length=128k, KV cache length=32k2025.1178.08—
DeepSeek-R1-0528Maximum sequence length=32k, KV cache length=32k2025.1173.97—
DeepSeek-R1-0528Maximum sequence length=128k, KV cache length=Full2025.1171.43—
REBALANCEBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0346.58,000
BaselineBackbone=DeepSeek-R1-Distill-Qwen-7B2026.03449,000
REBALANCEBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0322.511,000
BaselineBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0319.512,000