List your company

Mathematical Reasoning on MATH-500 (Accuracy, Avg., Drop ↓)

95.6AccuracySC@16
-3.82421.98847.873.612Jan 21, 2026Jan 27, 2026Feb 3, 2026Feb 10, 2026Feb 17, 2026Feb 24, 2026Mar 3, 2026

Evaluation results

MethodLinks
SC@16Model=Deepseek-R1-Distill-Qwen32B2026.0395.6——
PG@16Model=Deepseek-R1-Distill-Qwen32B2026.0395.6——
BF16 BaselineModel=Qwen3-4B, W-Bits=BF162026.0195.3370.63—
Avg@16Model=Deepseek-R1-Distill-Qwen32B2026.0394.3——
Reasoning-QATModel=Qwen3-4B, W-Bits=W4A4KV42026.0191.460.78-9.85
FlatQuantModel=Qwen3-4B, W-Bits=W4A4KV42026.0189.9358.28-12.35
BF16 BaselineModel=R1-1.5B, W-Bits=BF162026.0184.448.72—
PG@16Model=Qwen2.5-7B-Instruct2026.0380.8——
SC@16Model=Qwen2.5-7B-Instruct2026.0380.2——
BF16 BaselineModel=Qwen3-0.6B, W-Bits=BF162026.017441.1—
Avg@16Model=Qwen2.5-7B-Instruct2026.0374——
Reasoning-QATModel=R1-1.5B, W-Bits=W4A4KV42026.0173.241.31-7.41
FlatQuantModel=R1-1.5B, W-Bits=W4A4KV42026.0164.838.39-10.33
Reasoning-QATModel=Qwen3-0.6B, W-Bits=W4A4KV42026.0130.2721.44-19.66
FlatQuantModel=Qwen3-0.6B, W-Bits=W4A4KV42026.0121.6717.34-23.76
QuaRotModel=R1-1.5B, W-Bits=W4A4KV42026.011.22.11-46.61
QuaRotModel=Qwen3-0.6B, W-Bits=W4A4KV42026.0104.84-36.26