Claims

Across 17 frontier models, Stage 1 pedagogical-judgment scores are high (mean 0.893) while rank correlations with Stage 2 (ρ=0.24) and Stage 3 (ρ=0.21) are weak, with workflow scores topping out at 0.704

Claim · All claims
Evidence · 1 study · design r? · q2
Currency · newest study 2026 · checked against its sources 2026-10-20 (automated, 1 of 1 entry) · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims

Reading the evidence codes

q is the design tier, i the impact, n the sample, and the kind with its r the rigour judged within that kind: what each code means.

This claim: no standardized effect is recorded in observations/ for it yet, so no benchmark reading is possible.

q2 i? Model rankings reshuffle across the three benchmark stages: Stage 2 correlates weakly with Stage 1 (ρ=0.24) and Stage 3 (ρ=0.21), while Stages 1 and 3 correlate more closely (ρ=0.58), and the median gap between a model's best and worst stage rank is seven positions. → Zixin Chen 2026 q2 i? Stage 1 scores are high under the bounded-evidence contract (mean 0.893; range 0.799–0.947), while no model exceeds 0.704 on Stage 3 workflows. → Zixin Chen 2026

Evidence

Zixin Chen 2026

Zixin Chen, Peng Liu, Rui Sheng, Haobo Li, Jianhong Tu, Xiaodong Deng, Kashun Shum, Dayiheng Liu, Huamin Qu. (2026). Are Agents Ready to Teach? A Multi-Stage Benchmark for Real-World Teaching Workflows. Preprint. https://arxiv.org/abs/2605.14322 doi:10.48550/arxiv.2605.14322

q2 · i? · design · r?

Evaluation of 17 frontier models on the TEACHARENA leaderboard (Table 6), a benchmark evaluation rather than a human experiment. The article reports "mean 0.893" on Stage 1, weak rank correlations between Stage 2 and Stages 1 and 3, and a median best-to-worst stage-rank gap of seven positions; workflow scores top out at 0.704.

"Stage 1 scores are high under its bounded-evidence contract (mean 0.893; range 0.799–0.947), while stage-resolved reporting reveals substantial rank variation. The descriptive rank correlations are weak between Stage 2 and both Stage 1 (ρ= 0.24 ) and Stage 3 (ρ= 0.21 )"

Discussion

Page metadata
Field Value
type claim
id teacharena-stage-rank-reshuffling
title Across 17 frontier models, Stage 1 pedagogical-judgment scores are high (mean 0.893) while rank correlations with Stage 2 (ρ=0.24) and Stage 3 (ρ=0.21) are weak, with workflow scores topping out at 0.704
description Across 17 frontier models, Stage 1 pedagogical-judgment scores are high (mean 0.893) while rank correlations with Stage 2 (ρ=0.24) and Stage 3 (ρ=0.21) are weak, with workflow scores topping out at 0.704
status review
evidence_strength moderate
newest_evidence 2026
source_checked 2026-10-20
recheck_by 2029
generated by process:wiki-ingest on 2026-10-10
trust unverified — no one has confirmed this page's content
sources 1 — zixin-chen-2026

Frontmatter is the page's machine-readable record — see the Schema & Guide.