# Grading Handwritten Engineering Exams with Multimodal Large Language Models

arXiv: https://arxiv.org/abs/2601.00730 · HTML: https://mlsanity.com/paper/2601.00730/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 26 references that are in the map.
- First posted on arXiv 2 Jan 2026.

## Builds on (most cited first, 7 of 7)

- [Survey on Automated Short Answer Grading with Deep Learning: from Word Embeddings to Transformers](https://mlsanity.com/paper/2204.03503/) (2022)
- [Mathematical Language Processing: Automatic Grading and Feedback for Open Response Mathematical Questions](https://mlsanity.com/paper/1501.04346/) (2015)
- [DrawEduMath: Evaluating Vision Language Models with Expert-Annotated Students' Hand-Drawn Math Images](https://mlsanity.com/paper/2501.14877/) (2025)
- [Performance of the Pre-Trained Large Language Model GPT-4 on Automated Short Answer Grading](https://mlsanity.com/paper/2309.09338/) (2023)
- [Can Large Language Models Make the Grade? An Empirical Study Evaluating LLMs Ability to Mark Short Answer Questions in K-12 Education](https://mlsanity.com/paper/2405.02985/) (2024)
- [Evaluating GPT-4 at Grading Handwritten Solutions in Math Exams](https://mlsanity.com/paper/2411.05231/) (2024)
- [Automated Grading of Students' Handwritten Graphs: A Comparison of Meta-Learning and Vision-Large Language Models](https://mlsanity.com/paper/2507.03056/) (2025)

## Abstract (arXiv)

Handwritten STEM exams capture open-ended reasoning and diagrams, but manual grading is slow and difficult to scale. We present an end-to-end workflow for grading scanned handwritten engineering quizzes with multimodal large language models (LLMs) that preserves the standard exam process (A4 paper, unconstrained student handwriting). The lecturer provides only a handwritten reference solution (100%) and a short set of grading rules; the reference is converted into a text-only summary that conditions grading without exposing the reference scan. Reliability is achieved through a multi-stage design with a format/presence check to prevent grading blank answers, an ensemble of independent graders, supervisor aggregation, and rigid templates with deterministic validation to produce auditable, machine-parseable reports. We evaluate the frozen pipeline in a clean-room protocol on a held-out real course quiz in Slovenian, including hand-drawn circuit schematics. With state-of-the-art backends (GPT-5.2 and Gemini-3 Pro), the full pipeline achieves ≈8-point mean absolute difference to lecturer grades with low bias and an estimated manual-review trigger rate of ≈17% at Dmax=40. Ablations show that trivial prompting and removing the reference solution substantially degrade accuracy and introduce systematic over-grading, confirming that structured prompting and reference grounding are essential.