SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2605.13167 · 2026

GeoBuildBench: A Benchmark for Interactive and Executable Geometry Construction from Natural Language

Huishuai Zhang, Rui Yang, Jinwoong Kim

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 11 functions out of this paper's own repositories and ran 10 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
ooongs/GeoBuildBench canonical 10 of 11
FunctionStatusWhere it lives
a_to_cpx Ran ooongs/GeoBuildBench/src/core/geo_types.py
pointer only (licence: NONE) · get_code("157ccb58eb7a99cc")
cpx_to_a Ran ooongs/GeoBuildBench/src/core/geo_types.py
pointer only (licence: NONE) · get_code("666b801e4beb1eb3")
ensure_dir Ran ooongs/GeoBuildBench/src/utils.py
pointer only (licence: NONE) · get_code("cd13711a78eb95fe")
evaluate_math_expression Ran ooongs/GeoBuildBench/src/core/random_constr.py
pointer only (licence: NONE) · get_code("5de4717df9f3d151")
find_failed_problems Ran ooongs/GeoBuildBench/resume_benchmark.py
pointer only (licence: NONE) · get_code("85b8769cd25af23e")
find_incomplete_problems Ran ooongs/GeoBuildBench/resume_benchmark.py
pointer only (licence: NONE) · get_code("c296574d8c8a2fae")
get_completed_problems Ran ooongs/GeoBuildBench/resume_benchmark.py
pointer only (licence: NONE) · get_code("f9d0b660e2aac910")
interpolate Ran ooongs/GeoBuildBench/src/core/geo_types.py
pointer only (licence: NONE) · get_code("302aa95d6b92f423")
parse_trig_function Ran ooongs/GeoBuildBench/src/core/random_constr.py
pointer only (licence: NONE) · get_code("8799dfa5f983c72d")
resolve_path Ran ooongs/GeoBuildBench/src/utils.py
pointer only (licence: NONE) · get_code("2e2818788a6e2c14")
get_output_dir Not yet run ooongs/GeoBuildBench/src/utils.py
pointer only (licence: NONE) · get_code("9b0c75ec4659dbad")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We introduce GeoBuildBench, a benchmark designed to evaluate whether large language models and multimodal agents can ground informal natural-language plane geometry problems into executable geometric constructions. Unlike existing geometry benchmarks that focus on answer correctness or static diagram interpretation, GeoBuildBench treats geometry diagram as an interactive construction task: given a textual problem, an agent must generate a domain-specific language (DSL) program to produce a diagram satisfying explicitly specified geometric objects and verifiable constraints. The benchmark features 489 Chinese textbook-style problems, curated through automated filtering and human validation to ensure text-complete, constructible problem specifications. We evaluate several state-ofthe-art multimodal models in a bounded iterative setting and show that, despite reasonable success rates, models frequently exhibit structural hallucinations, missing objects, and failures to satisfy geometric constraints, with limited ability to exploit visual and constraintbased feedback for self-correction. These results highlight geometry construction as a rigorous testbed for grounded, executable reasoning beyond textual or visual plausibility. Our benchmark and code are released at https: //github.com/ooongs/GeoBuildBench.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2605.13167")
get_code_for_paper("2605.13167")
have("2605.13167")

Connect an agent — have() is free.