List your company

Agentic Task Execution on CAPTRACEBENCH Easy 1.0

81.4Success Rate (SR)Claude-Opus-4.6
64.65669.00373.3577.697Jun 9, 2026

Evaluation results

MethodLinks
Claude-Opus-4.6Setting=w/ Oracle Skills2026.0681.491.1
Claude-Opus-4.6Setting=Raw Traces (w/ Extracted Skills)2026.068087
Claude-Opus-4.6Setting=Raw Traces (w/ Evolved Skills)2026.0674.586
Claude-Opus-4.6Setting=Raw Traces (w/ Retrieval Reuse)2026.0673.886
Claude-Opus-4.6Setting=REDACT Protected (w/ Extracted Skills)2026.0672.580.9
Gemini-3-ProSetting=No Skills2026.0672.383.6
GPT-5.2-CodexSetting=No Skills2026.0671.386.7
Claude-Sonnet-4.6Setting=No Skills2026.0670.580.6
Average.Setting=No Skills2026.0669.682.9
Average.Setting=REDACT Protected (w/ Extracted Skills)2026.0669.380
Gemini-3-FlashSetting=No Skills2026.0669.283.4
Claude-Opus-4.6Setting=No Skills2026.0668.884
Claude-Opus-4.6Setting=REDACT Protected (w/ Evolved Skills)2026.0667.682.1
Claude-Opus-4.6Setting=REDACT Protected (w/ Retrieval Reuse)2026.0666.881.7
Claude-Haiku-4.5Setting=No Skills2026.0665.379.1