List your company

Search-based Question Answering on SearchQA

42Hotpot ScoreHarnessForge
28.136831.735935.33538.9341Jun 1, 2026

Evaluation results

MethodLinks
HarnessForgeBackbone=Qwen3-4B2026.06424242
GRPOBackbone=Qwen3-8B2026.0641.334241.67
HarnessForgeBackbone=Qwen3-8B2026.0641.3344.3342.83
SFTBackbone=Qwen3-8B2026.0640.6741.3341
RLOOBackbone=Qwen3-8B2026.064039.3339.67
SFTBackbone=Qwen3-4B2026.0636.6738.6737.67
GRPOBackbone=Qwen3-4B2026.063643.3339.67
RLOOBackbone=Qwen3-4B2026.0635.334037.67
MermaidFlowBackbone=Qwen3-8B2026.0632.6730.6731.67
MaASBackbone=Qwen3-8B2026.06323232
MermaidFlowBackbone=Qwen3-4B2026.0631.332829.67
ADASBackbone=Qwen3-8B2026.0631.333231.67
AFlowBackbone=Qwen3-8B2026.0631.3332.6732
AFlowBackbone=Qwen3-4B2026.0630.6729.3330
MaASBackbone=Qwen3-4B2026.0630.673030.33
AgentSquareBackbone=Qwen3-4B2026.0629.333029.67
AgentSquareBackbone=Qwen3-8B2026.0629.333029.5
ADASBackbone=Qwen3-4B2026.0628.672828.33