benchgap
Calibration

GDP.pdf → AA-AnalystAgent

AA-AnalystAgent is estimated from GDP.pdf with a Hill curve fitted on 10 models measured on both: y = 0.0491 + (0.5473 − 0.0491)·x^4.81 / (0.14199^4.81 + x^4.81), R² = 0.94, cross-validated error 6.5 pp. It is used for 4 estimates.

Estimated modelGDP.pdfAA-AnalystAgentSource
GPT-6 Luna22.8%50.1%estimated ± 6.5 pp, medium confidence
Grok 4.720.0%46.7%estimated ± 6.5 pp, medium confidence
MiMo-V2.6-Pro19.2%45.3%estimated ± 6.5 pp, medium confidence
Mistral Large 418.6%44.0%estimated ± 6.5 pp, medium confidence