List your company

Mathematical Reasoning on AIME 2023 (Accuracy and Token Count)

73.33Accuracy (%)RAG
-2.933216.865936.66556.4641Sep 26, 2025Nov 2, 2025Dec 10, 2025Jan 17, 2026Feb 24, 2026Apr 3, 2026May 11, 2026

Evaluation results

MethodLinks
RAGBackbone=Qwen3-14B2025.0973.3311,750.4
CoT-SCBackbone=Qwen3-4B2025.097044,736.6
CoTBackbone=Qwen3-4B2025.0963.3311,366.2
CoT-SCBackbone=Qwen3-8B2025.0963.3349,529.87
CoT-SCBackbone=Qwen3-14B2025.0963.3346,998.97
RoT + TIBackbone=Qwen3-14B2025.0963.3310,623.03
RAGBackbone=Qwen3-8B2025.096012,728.53
CoTBackbone=Qwen3-14B2025.096012,096.93
RoTBackbone=Qwen3-14B2025.0956.6712,430.8
RAGBackbone=Qwen3-4B2025.0953.3311,829.9
RoTBackbone=Qwen3-8B2025.0953.3313,341.63
CoTBackbone=Qwen3-8B2025.095012,812.63
BoTBackbone=Qwen3-8B2025.095056,428.9
RoT + TIBackbone=Qwen3-8B2025.095010,698.93
RoTBackbone=Qwen3-4B2025.0946.6713,277.83
RoT + TIBackbone=Qwen3-4B2025.0946.6710,003.3
BoTBackbone=Qwen3-4B2025.0936.449,331.67
CoTBackbone=Qwen3-1.7B2025.0933.3311,676.13
RAGBackbone=Qwen3-1.7B2025.0933.3311,456.13
CoT-SCBackbone=Qwen3-1.7B2025.0933.3345,657.83
RoTBackbone=Qwen3-1.7B2025.0933.3312,577.16
EVOCHAMBERBackbone=Qwen3-8B, Team size (k)=32026.0533.3—
BoTBackbone=Qwen3-1.7B2025.0923.3345,898.9
AgentNetBackbone=Qwen3-8B, Team size (k)=32026.0516.7—
MemCollabBackbone=Qwen3-8B, Team size (k)=32026.0516.7—
CoT-SCBackbone=Qwen3-0.6B2025.0916.6742,095.83
RoT + TIBackbone=Qwen3-1.7B2025.0916.678,292.7
CoTBackbone=Qwen3-0.6B2025.0913.3310,421.53
SC (k=5)Backbone=Qwen3-8B, Team size (k)=52026.0513.3—
EvoMemBackbone=Qwen3-8B, Team size (k)=32026.0513.3—
BoTBackbone=Qwen3-0.6B2025.091330,258.34
RAGBackbone=Qwen3-0.6B2025.091010,849.5
RoTBackbone=Qwen3-0.6B2025.091010,933.53
SABackbone=Qwen3-8B, Team size (k)=12026.0510—
DyLANBackbone=Qwen3-8B, Team size (k)=32026.056.7—
RoT + TIBackbone=Qwen3-0.6B2025.0903,058.8