K12-Train-Full achieves the best overall performance on three multimodal educational benchmarks, and textual and visual supervision are complementary¶

Claim · All claims
Evidence · 1 study (2 entries) · design r2 · 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¶

q2 i? K12-Train-Full achieves the best overall result on Gaokao-MM, MDK12-Bench, and K12Vista, consistently outperforming both Text-Only and MM-Only configurations, indicating the two supervision sources are complementary. → Hao Liang 2026 q2 i? On K12Vista, K12-Train-Full is the only fine-tuned configuration that improves upon the untuned base model (79.95 vs. 79.72). → Hao Liang 2026 (2)

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

LoRA SFT of Qwen3.5-2B-Base on K12-Train-Full, K12-Train-Text, K12-Train-MM, and full DataFlow and WizardLM datasets, evaluated on Gaokao-MM, MDK12-Bench, and K12Vista (Tables 6-8). The article reports K12-Train-Full surpassing the full DataFlow and WizardLM baselines with a smaller sample size (7,335 vs. 10,000/142,759).

"K12-Train-Full achieves the best overall result on all three multimodal benchmarks and consistently outperforms both Text Only and MM Only, indicating that the two supervision sources are complementary."

Hao Liang 2026 (2)¶

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

q2 · i? · design · r2

K12Vista evaluation across five subjects and three question formats (Table 8), described as a rigorous test of capability retention since the untuned base model already scores 79.72. K12-Train-Full achieved the best results across multiple-choice, fill-in-the-blank, and open-ended QA formats.

"K12-Train-Full is the only fine-tuned configuration that improves upon the base model, reaching the best average score of79.95."

Discussion¶

Related Claims¶

• Epi2Diff achieves the best performance across all four item-difficulty test sets and all reported metrics, outperforming SLM fine-tuning, LLM prompting, and supervised LLM adaptation baselines — related
• SFT on K12-Train improves out-of-domain subjects (Chinese, Humanities Math), suggesting acquisition of a transferable structurally-grounded answer style — related
• Under a matched 2,300-sample SFT budget, K12-Train-Text outperforms eight mainstream instruction-tuning corpora on GaokaoBench and EduEval across two base models — related
• 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 — related
• Education-oriented tuning improves authentic K-12 problem solving, with OmniEdu-27B's MDK12-Bench open-question correctness rising from 62.08% to 78.77% — related
• OmniEdu-27B achieves the best overall curriculum-grounding performance among evaluated models, raising K12-Bench EM from 52.11% to 63.12% — related
• Education-oriented instruction tuning consistently improves curriculum grounding, K-12 problem solving, and pedagogical tutoring at every evaluated model scale — related
• Education-oriented tuning produces the largest gains in scaffolded instruction and use of longitudinal student evidence, with LongTutor Evidence averages rising sharply at all scales — related