List your company
ResearchFormerly WizWand

Mathematical Reasoning on AMO-Bench

11.8Mean@64 AccuracyAgentic Proposing
1.6084.2546.99.546Feb 1, 2026Feb 2, 2026Feb 3, 2026
Updated 8mo ago

Evaluation results

MethodLinks
Agentic Proposingtraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.0211.8
Deepmathtraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.0210.3
PromptCot 2.0training budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.0210.2
Polaristraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.029.8
Qwen3-4B-Instruct-2507mode=zero-shot2026.029.3
PromptCoTtraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.029.1
MathSmithtraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.029.1
Socratic-zerotraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.029.1
OpenR1mathtraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.028.9
DeepSeek-V3.2-Spetraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.028.8
Qwen3-Maxtraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.028.5
OpenthoughtsS3training budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.028.4
NuminaMathtraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.028.2
Gemini-3-Protraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.028.1
Metamathtraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.027.8
GPT-5.2-Hightraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.027.8
Wizardmathtraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.027.6
Claude4.5-Opustraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.027.4
R-zerotraining budget=10,000-trajectory, recipe=GRPO, backbone=Qwen3-4B-Instruct-25072026.026.7
DynamicBackbone=Llama-3.1-8B2026.024
Dynamic + PRM SelectionBackbone=Llama-3.1-8B2026.024
DynamicBackbone=Qwen-2.5-7B2026.024
Dynamic + PRM SelectionBackbone=Qwen-2.5-7B2026.024
DirectBackbone=Llama-3.1-8B2026.022
Direct + PRM SelectionBackbone=Llama-3.1-8B2026.022
DirectBackbone=Qwen-2.5-7B2026.022
Direct + PRM SelectionBackbone=Qwen-2.5-7B2026.022