List your company

Mathematical Reasoning on Mathematical Reasoning Benchmarks Average

78.1AccuracyQwen3-8B (Thinking)
10.39627.97345.5563.127May 16, 2026

Evaluation results

MethodLinks
Qwen3-8B (Thinking)Model=Qwen3-8B, Variant=Thinking2026.0578.187.3—
Qwen3-1.7B (Thinking)Model=Qwen3-1.7B, Variant=Thinking2026.0564.180.3—
Qwen3-1.7B + Guided (ρ¯ ≈ 0.16)Model=Qwen3-1.7B, Intervention Rate (rho)=≈ 0.162026.0562.183.8112
Qwen3-0.6B + Guided (ρ¯ ≈ 0.13)Model=Qwen3-0.6B, Intervention Rate (rho)=≈ 0.132026.0552.480157
Qwen3-1.7B + Guided (ρ¯ ≈ 0.06)Model=Qwen3-1.7B, Intervention Rate (rho)=≈ 0.062026.0543.673.576
Qwen3-0.6B (Thinking)Model=Qwen3-0.6B, Variant=Thinking2026.0543.464.1—
Qwen3-1.7B-BaseModel=Qwen3-1.7B, Variant=Base2026.0530.851.8—
Qwen3-0.6B + Guided (ρ¯ ≈ 0.04)Model=Qwen3-0.6B, Intervention Rate (rho)=≈ 0.042026.0529.161.491
Qwen3-0.6B-BaseModel=Qwen3-0.6B, Variant=Base2026.051336—