List your company

Task Completion on SOP-Bench

100AccuracyGA
79.284.69095.4Jun 6, 2026

Evaluation results

MethodLinks
GAModel=Claude Sonnet 4.62026.061002.02532.080.48
OpenClawModel=Claude Sonnet 4.62026.061002.6402.640.38
GAModel=deepseek-v4-pro2026.061008574590222.18
BA-FullModel=deepseek-v4-pro2026.061009145196520.73
BA-IncModel=deepseek-v4-pro2026.06100000—
BA-FullModel=deepseek-v4-flash2026.06951.2601.2615.13
BA-IncModel=deepseek-v4-flash2026.0695145815319.63
Claude CodeModel=Claude Sonnet 4.62026.06851.23231.250.68
GAModel=deepseek-v4-flash2026.06801.34571.3911.47