List your company

Mathematical Reasoning on AMC23 (Accuracy, Latency, Bit Metrics)

68.3AccuracyBS (B)
9.43624.7184055.282Jun 15, 2025Jul 2, 2025Jul 19, 2025Aug 5, 2025Aug 22, 2025Sep 8, 2025Sep 26, 2025

Evaluation results

MethodLinks
BS (B)Model Family=Qwen2.5 Family, n=162025.0668.3352.8100
SPECSModel Family=Qwen2.5 Family, n=162025.0668.328.82.770.3
SPECSModel Family=Qwen2.5 Family, n=82025.0662.520.71.9462.8
BS (B)Model Family=Qwen2.5 Family, n=42025.0659.218.31.46100
SPECSModel Family=Qwen2.5 Family, n=42025.0659.214.81.4267.9
BS (B)Model Family=Qwen2.5 Family, n=82025.0659.221.62100
BS (S)Model Family=Qwen2.5 Family, n=162025.0650.825.61.810
RSD++Model Family=Qwen2.5 Family, n=162025.0650252.5655.6
RSD++Model Family=Qwen2.5 Family, n=82025.0649.220.31.9961.1
BS (S)Model Family=Qwen2.5 Family, n=82025.0646.717.61.140
BS (S)Model Family=Qwen2.5 Family, n=42025.0638.310.30.80
M_RLP + PostPre-training Condition=Reinforcement Learning from Pretraining, Post-training=SFT + RLVR2025.0936.5———
SPECSModel Family=Llama3 Family, n=162025.0636.344.21.5698.1
M_base + PostPre-training Condition=Base, Post-training=SFT + RLVR2025.0935———
M_CPT + PostPre-training Condition=Continuous Pretraining, Post-training=SFT + RLVR2025.0934.38———
BS (B)Model Family=Llama3 Family, n=162025.0632.532.11.24100
M_CPTPre-training Condition=Continuous Pretraining, Post-training=None2025.0931.25———
M_RLPPre-training Condition=Reinforcement Learning from Pretraining, Post-training=None2025.0931.25———
RSD++Model Family=Qwen2.5 Family, n=42025.0631.212.60.8260.5
SPECSModel Family=Llama3 Family, n=82025.0630281.0493.6
BS (B)Model Family=Llama3 Family, n=82025.0627.519.10.84100
BS (B)Model Family=Llama3 Family, n=42025.0626.312.70.59100
SPECSModel Family=Llama3 Family, n=42025.0626.315.40.6596.4
M_basePre-training Condition=Base, Post-training=None2025.0925.94———
BS (S)Model Family=Llama3 Family, n=162025.0624.224.471.130
RSD++Model Family=Llama3 Family, n=162025.062016.31.1890
RSD++Model Family=Llama3 Family, n=82025.0618.310.80.9190
RSD++Model Family=Llama3 Family, n=42025.06158.90.6691.8
BS (S)Model Family=Llama3 Family, n=82025.0614.211.850.610
BS (S)Model Family=Llama3 Family, n=42025.0611.75.940.380