List your company

Web-based Question Answering on WebQA (200 held-out questions)

62.7AccuracyMAGE
33.37240.98648.656.214May 11, 2026

Evaluation results

MethodLinks
MAGEBackbone=Qwen3-8B, Judge=Claude Sonnet 4.6, MAGE Executor Model=Opus2026.0562.77.7
MAGEBackbone=Qwen3-8B, Judge=Claude Sonnet 4.6, MAGE Executor Model=Sonnet2026.0559.67.7
0-shot CoTBackbone=Qwen3-8B, Judge=Claude Sonnet 4.6, Prompting Scaffold=0-shot CoT2026.0555—
SC10Backbone=Qwen3-8B, Judge=Claude Sonnet 4.6, Prompting Scaffold=SC10, Samples=102026.0546.5—
8-shotBackbone=Qwen3-8B, Judge=Claude Sonnet 4.6, Prompting Scaffold=8-shot2026.0542—
ReActBackbone=Qwen3-8B, Judge=Claude Sonnet 4.6, Prompting Scaffold=ReAct2026.0535.5—
ReflexionBackbone=Qwen3-8B, Judge=Claude Sonnet 4.6, Prompting Scaffold=Reflexion2026.0534.5—