List your company

Mathematical Reasoning on MATH 500 (Acc, Delta_Up, Delta_Down)

89.8AccuracyAgentic-Verifier-Qwen3-4B
53.19262.69672.281.704Apr 17, 2026

Evaluation results

MethodLinks
Agentic-Verifier-Qwen3-4BRefinement Turn=32026.0489.81.81.2
Agentic-Verifier-Qwen3-4BRefinement Turn=22026.0489.22.41.2
DS-Distill-14BRefinement Turn=22026.0487.45.81.4
DS-Distill-14BRefinement Turn=32026.0485.82.23.8
Qwen3-4B-ThinkRefinement Turn=32026.0484.82.42.2
Qwen3-4B-ThinkRefinement Turn=22026.0484.661.4
Agentic-Verifier-Qwen3-4BRefinement Turn=12026.0484.241.60.6
DS-Distill-14BRefinement Turn=12026.048337.41.4
Qwen3-4B-ThinkRefinement Turn=12026.048033.80.8
Mistral-Small-24B-InstructRefinement Turn=32026.0467.84.43
Mistral-Small-24B-InstructRefinement Turn=22026.0466.45.80.8
Qwen2.5-7B-InstructRefinement Turn=32026.04664.23
Qwen2.5-7B-InstructRefinement Turn=22026.0464.86.21.8
Mistral-Small-24B-InstructRefinement Turn=12026.0461.415.61.2
Qwen2.5-7B-InstructRefinement Turn=12026.0460.4140.6
Llama3.1-8B-InstructRefinement Turn=32026.0458.843.6
Llama3.1-8B-InstructRefinement Turn=22026.0458.462.2
Llama3.1-8B-InstructRefinement Turn=12026.0454.6113.4