Benchmarking PhD-Level Coding in 3D Geometric Computer Vision

TL;DR

GeoCodeBench is a new benchmark for evaluating AI models' ability to generate correct 3D geometric vision code, revealing current models' limited performance and highlighting challenges in scientific coding comprehension.

Contribution

The paper introduces GeoCodeBench, a comprehensive benchmark for PhD-level 3D vision coding tasks, with a novel evaluation framework and insights into model capabilities.

Findings

GPT-5 achieves only 36.6% pass rate on the benchmark.

Research tasks are significantly harder than general 3D capability tasks.

Shorter context inputs (up to Method section) outperform full-paper inputs.

Abstract

AI-assisted coding has rapidly reshaped software practice and research workflows, yet today's models still struggle to produce correct code for complex 3D geometric vision. If models could reliably write such code, the research of our community would change substantially. To measure progress toward that goal, we introduce GeoCodeBench, a PhD-level benchmark that evaluates coding for 3D vision. Each problem is a fill-in-the-function implementation task curated from representative papers at recent venues: we first let a tool propose candidate functions from official repositories, then perform careful human screening to select core 3D geometric components. For every target, we generate diverse, edge-case unit tests, enabling fully automatic, reproducible scoring. We evaluate eight representative open- and closed-source models to reflect the current ecosystem. The best model, GPT-5, attains only 36.6% pass rate, revealing a large gap between current capabilities and dependable 3D scientific coding. GeoCodeBench organizes tasks into a two-level hierarchy: General 3D capability (geometric transformations and mechanics/optics formulation) and Research capability (novel algorithm implementation and geometric logic routing). Scores are positively correlated across these axes, but research-oriented tasks are markedly harder. Context ablations further show that "more paper text" is not always better: cutting off at the Method section statistically outperforms full-paper inputs, highlighting unresolved challenges in long-context scientific comprehension. Together, these findings position GeoCodeBench as a rigorous testbed for advancing from generic coding to trustworthy 3D geometric vision coding.