Current LLMs lack robust curriculum cognition: even the best model reaches only 57.1% exact match on K12-Bench, with prerequisite and neighbor tasks hardest¶
Claim · All claims
Evidence · 1 study · designr2·q2
Currency · newest study 2026 · not yet checked against its sources · 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? On K12-Bench, the best overall exact match is 57.1% (Gemini-3-Flash) and a strong open-source model (Gemma-4-31B-IT) reaches 46.4%, with Prereq and Neighbor yielding the lowest EM across almost every model. → Hao Liang 2026
Evidence¶
Hao Liang 2026¶
Hao Liang, Qihan Lin, Mingrui Chen, Hengyi Feng, Zhaoyang Han, Xiaochen Ma, Zhen Hao Wong, Meiyi Qiang, Linzhuang Sun, Wentao Zhang. (2026). K12-KGraph: A Curriculum-Aligned Knowledge Graph for Benchmarking and Training Educational LLMs. arXiv preprint. https://arxiv.org/abs/2605.09635 doi:10.48550/arxiv.2605.09635
q2 · i? · design · r2
Zero-shot, answer-only evaluation of five open-source LLMs and five proprietary APIs on K12-Bench, reporting per-task exact match and instance-level macro F1 (Table 3). The article reports "The best overall exact match is only 57.1%" for Gemini-3-Flash and 46.4% for Gemma-4-31B-IT. LLaMA-3-8B-Instruct (EM 7.2%) was essentially indistinguishable from random guessing (EM 6.7%).
"The best overall exact match is only 57.1% (Gemini-3-Flash); a strong open-source model (Gemma-4-31B-IT) reaches just 46.4%. In more than half the questions, models fail to identify the complete correct answer set."
Discussion¶
Related Claims¶
- K12-Train-Full achieves the best overall performance on three multimodal educational benchmarks, and textual and visual supervision are complementary — related
- OmniEdu-27B achieves the best overall curriculum-grounding performance among evaluated models, raising K12-Bench EM from 52.11% to 63.12% — reports the opposite
- Education-oriented instruction tuning consistently improves curriculum grounding, K-12 problem solving, and pedagogical tutoring at every evaluated model scale — related
Page metadata
| Field | Value |
|---|---|
type |
claim |
id |
llms-lack-curriculum-cognition-k12-bench |
title |
Current LLMs lack robust curriculum cognition: even the best model reaches only 57.1% exact match on K12-Bench, with prerequisite and neighbor tasks hardest |
description |
Current LLMs lack robust curriculum cognition: even the best model reaches only 57.1% exact match on K12-Bench, with prerequisite and neighbor tasks hardest |
status |
review |
evidence_strength |
moderate |
newest_evidence |
2026 |
recheck_by |
2029 |
generated |
by process:wiki-ingest on 2026-10-11 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — hao-liang-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.