List your company
ResearchFormerly WizWand

Mathematical Reasoning on MATH (Performance and Cost Metrics)

69.34AccuracyFlowBank
53.79257.828561.86565.9015Jun 9, 2026
Updated 4mo ago

Evaluation results

MethodLinks
FlowBankExecutor LLM=GPT-4o mini, Optimizer LLM=Qwen3-8B2026.0669.341.78
MaASExecutor LLM=GPT-4o mini2026.0665.022.43
AFlow (GPT-4o)Executor LLM=GPT-4o mini, Optimizer LLM=GPT-4o2026.0663.992.77
AFlow (Qwen3-8B)Executor LLM=GPT-4o mini, Optimizer LLM=Qwen3-8B2026.0663.243.23
AgentSquareExecutor LLM=GPT-4o mini, Optimizer LLM=GPT-4o2026.0662.761.35
ScoreFlowExecutor LLM=GPT-4o mini, Generator LLM=Qwen3-8B2026.06622.36
Multi-agent DebateExecutor LLM=GPT-4o mini2026.0658.927.66
Self-ConsistencyExecutor LLM=GPT-4o mini2026.0657.823.38
MultiPersonaExecutor LLM=GPT-4o mini2026.0657.410.56
MedPromptExecutor LLM=GPT-4o mini2026.0656.695.64
IOExecutor LLM=GPT-4o mini2026.0656.580.51
ADASExecutor LLM=GPT-4o mini, Optimizer LLM=GPT-4o2026.0656.172.95
CoTExecutor LLM=GPT-4o mini2026.0655.760.53
Self-RefineExecutor LLM=GPT-4o mini2026.0655.141.08
GPTSwarmExecutor LLM=GPT-4o mini2026.0654.941.69
ComplexCoTExecutor LLM=GPT-4o mini2026.0654.390.62