List your company

Mathematical Reasoning on GSM8K (Solve Rate)

94.4Solve RateAGENTCO-OP
30.991247.453163.91580.3769May 19, 2026May 26, 2026Jun 3, 2026Jun 10, 2026Jun 18, 2026Jun 25, 2026Jul 3, 2026

Evaluation results

MethodLinks
AGENTCO-OPBackbone=GPT-4o-mini2026.0594.4
ReConcileBackbone=GPT-4o-mini2026.0593.7
AFlow*Backbone=Mixed2026.0593.5
MultiPersonaBackbone=GPT-4o-mini2026.0592.8
IOBackbone=GPT-4o-mini2026.0592.7
CoT SCBackbone=GPT-4o-mini, Prompting Setting=5-shot2026.0592.7
CoTBackbone=GPT-4o-mini2026.0592.4
LLM-DebateBackbone=GPT-4o-mini2026.0592.4
ADASBackbone=GPT-4o-mini2026.0590.8
MedPromptBackbone=GPT-4o-mini2026.0590
Self RefineBackbone=GPT-4o-mini2026.0589.6
AFlowBackbone=GPT-4o-mini2026.0586.8
Olmo3 BaseTuning Protocol=Freezing param.2026.0737
LMK token tuningTuning Protocol=Freezing param.2026.0737
HiLS-Attn HoPE-Q-CalTuning Protocol=Full param. CPT2026.0736.85
HiLS-Attn RoPE-Q-CalTuning Protocol=Full param. CPT2026.0735.71
HiLS-Attn NoPE-Q-CalTuning Protocol=Full param. CPT2026.0735.03
Olmo3 512swaTuning Protocol=Full param. CPT2026.0733.43