Skip to content

ReFigBench: Benchmarking Scientific Figure Reconstruction as Editable PowerPoint Artifacts

Sep 2026 · 0 citations · 25 references
Computer Science

TL;DR

ReFigBench, a benchmark and evaluation framework built on 1,000 real overview figures retrieved from arXiv papers with full provenance, is introduced, exposing the tension between fidelity and editability as the central challenge for practical multimodal document agents.

Abstract

Multimodal coding agents are expected to turn visual inputs into usable artifacts, and they act through a harness, the layer of tools, context management, and execution environment around the model. Existing evaluations often isolate short tool calls, API traces, or screenshot resemblance, and a low score under these proxies cannot say whether the model saw poorly, planned poorly, or was failed by its harness. We study scientific overview figure reconstruction, an agent task in which a source image must become an editable PowerPoint slide that preserves text, topology, layout, and native document structure. We introduce ReFigBench, a benchmark and evaluation framework built on 1,000 real overview figures retrieved from arXiv papers with full provenance. Coding agents from four model families reconstruct every figure under two workflows, direct code generation and a specialized PPTX workflow, and the strongest model runs inside two commercial harnesses, yielding ten configurations. Evaluation combines deterministic artifact checks, repeated automated scoring by judges from two model families, and blinded human comparisons. Perception remains a bottleneck that iterative rendering only partly repays. Whether workflow effort converts into quality depends on the model together with its harness, since the same model gains from the specialized workflow inside one harness and loses inside the other, and the harness shifts scores even under an identical direct prompt. The specialized workflow erases native connectors in every configuration, human judges still prefer its renderings in most matchups, and even the strongest agent falls short of the rubric ceiling. These results expose the tension between fidelity and editability as the central challenge for practical multimodal document agents.

View source

Similar papers

Review Sep 2026

SciFigure2Code: An AI-Reconstructed Benchmark for Scientific Figure-to-Code

SciFigure2Code is introduced, an AI-reconstructed benchmark that instead evaluates presentation recovery: generating editable Python programs that preserve how a scientific panel is arranged and read, and provides an auditable testbed for agents that construct editable, visually faithful scientific figure presentations...

Wen-Tao Li, Yi-Bo Wu, Yi-Zhe Chen et al. · 0 citations

Code4Scene: Benchmarking Coding Agents for Constructing and Editing 3D Scenes

This work introduces Code4Scene, a benchmark of 190 Unreal Engine cases built from human-assembled scenes that evaluates coding agents on two complementary settings under a shared execution interface, and exposes a gap between plausible 3D generation and reliable spatial reasoning and state control.

Xiao-Kang Ye, Siddhant Hitesh Mantri, Zi-Meng Chen et al. · 3 citations
#artificial intelligence Preprint Aug 2026

PPTBench: Can Coding Agents Reconstruct the Visual World through Structured, Editable Slides

PPTBench establishes a measurable testbed for studying visual coding and advancing agents toward more reliable visual creation, and shows that agents can generally produce valid slide files, but struggle to produce high-quality reconstructions that faithfully recover the semantics and visual structure of the target.

Xiao-Qiu Wang, Yi-Zhe Chi, Wen-Yi Li et al. · 0 citations
Preprint Sep 2026

Omni2Web: Benchmarking Audiovisual Website Development

Screen-recorded web editing requests contain weak deictic expressions such as ``this''and ``there,''whose referents depend on speech, cursor trajectories, page state, and edit history. Such requests require intent recovery beyond the explicit specifications assumed by many existing web-editing benchmarks. We introduce...

Ming-Hao Han, Zhenghao Xing, Xi-Ze Cheng et al. · 1 citation · ⚡1
Preprint Aug 2026

VisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback?

This work introduces VisEditBench, a benchmark of 1,395 human-annotated visualization code-editing tasks grounded in realistic visualization workflows and failure cases, and proposes VisEditAgent, a render-grounded editing framework that iteratively generates, executes, validates, and refines candidate edits.

Mizanur Rahman, Arshia Azimlu, Shadikur Rahman et al. · 3 citations · ⚡1
Conference Aug 2026

Harness Engineering for Multi-Agent Data Visualization with Small Language Models

Turning a natural-language question into a correct, publication-ready data visualization usually takes several rounds of coding, inspection, and debugging. Large language models can draft plotting code, but a single-shot model call has no way to run the code, look at the resulting image, recover from execution errors,...

Pei-Lin Wang, Xin-Xiao Li, Eisei Nakahara · 0 citations

Related blog posts

MIT News · Artificial Intelligence Jun 3, 2026

MIT researchers teach AI models to interpret charts

The new ChartNet training dataset could improve the accuracy of vision-language models that help analyze business trends or interpret scientific figures.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.