Estimated human baseline on the CDPK benchmark is approximately 50%, based on aggregate Chilean teacher exam results¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person · recheck by 2030: baselines and conditions change

Subclaims¶

q2 i? Based on results of more than 25,000 teachers from 2017 to 2021, the mean expected score of a human trainee teacher on the CDPK benchmark is estimated at approximately 50%. → Lelièvre 2025

Evidence¶

Lelièvre 2025¶

Lelièvre, M., Waldock, A., Liu, M., Valdes Aspillaga, N., Mackintosh, A., Ogando Portela, M. J., Lee, J., Atherton, P., Ince, R. A. A., & Garrod, O. G. B. (2025). Benchmarking the Pedagogical Knowledge of Large Language Models. https://arxiv.org/abs/2506.18710 doi:10.48550/arxiv.2506.18710

q2 · i? · design · r2

Estimation from aggregate subject- and year-level results of Chilean teachers on ECEP exams, allocating each benchmark question the average accuracy of its source paper; four methods all produced estimates around 50%. Question-level human results were not available, so this is an estimate only.

"Based on the results of more than 25,000 teachers from 2017 to 2021, the mean expected score of a human trainee teacher would be expected to be approximately 50% on The Pedagogy Benchmark."

Discussion¶

Related Claims¶

• The benchmark procedure is stable under stochastic decoding, with a maximum accuracy standard deviation of 0.43% over 20 repeated runs — related
• Accuracy-cost value frontier: at $0.10 per million tokens, CDPK performance rose from 50% (April 2024) to 82% (June 2025) — related
• SEND and CDPK benchmark results correlate highly (r = 0.94), but higher-performing models do relatively better on general pedagogy than SEND — related
• LLM accuracy on the CDPK pedagogy benchmark ranges from 28% to 89%, with closed-weight reasoning models dominating the top of the leaderboard — related