# huggingface/Math-Verify

Repository: https://github.com/huggingface/Math-Verify
Canonical: https://ross.abutalabs.com/products/math-verify
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-01-10T01:49:27+00:00

## Health v2 (maintenance only)
Score: 50/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 61, release rhythm 40, longevity 42
- inputs: {"age_days": 593, "days_push": 236, "days_rel": 427, "gap_med": 0, "n_releases_24m": 10}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1186, forks 58 (observed 2026-08-28T04:03:55.182831+00:00)

## What it is
A Python library from Hugging Face for robustly parsing and verifying mathematical expressions, designed to evaluate Large Language Model outputs on math tasks. It extracts answers from LaTeX, plain expressions, and strings, then checks equivalence with flexible comparison logic.

## Use cases
- evaluate llm outputs on math benchmarks like MATH
- verify if two mathematical expressions are equivalent
- extract answers from latex model responses
- grade math answers in an evaluation harness
- parse latex math expressions in python
- compare set and interval expressions for equality

## When to choose
- you are benchmarking LLMs on mathematical reasoning tasks
- you need robust extraction of answers from free-form or LaTeX model output
- existing evaluators like the Harness or Qwen's undercount model performance
- you need exact handling of sets, intervals, equations, and units

## When to avoid
- you need a general-purpose computer algebra system
- you only need simple numeric comparison without answer extraction
- your project is not Python-based

## Facets
- artifact type: library
- maturity: active
- function: parser, nlp, testing, math
- domain: large-language-models, education
- platform: python
- tags: math-evaluation, llm-benchmarking, latex-parsing, answer-verification, antlr, natural-language-processing, algorithms

## Member repositories
- huggingface/Math-Verify (main) score 50

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:55.182831+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:23:53.697971+00:00, confidence not recorded.
  - readme: https://github.com/huggingface/Math-Verify (fetched 2026-08-28T04:03:55.182831+00:00, sha e0fca6ffcd4f)
  - registry_pypi: https://pypi.org/pypi/math-verify/json (fetched 2026-08-29T12:31:05.458252+00:00, sha b5c503a11f73)
- Data as of 2026-08-30T08:39:29.467469+00:00.