# langchain-ai/openevals

Readymade evaluators for your LLM apps

Repository: https://github.com/langchain-ai/openevals
Canonical: https://ross.abutalabs.com/products/openevals
Language: Python
License: MIT
License Family: permissive
Last push: 2026-08-20T07:55:05+00:00

## Health v2 (maintenance only)
Score: 86/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 98, longevity 40
- inputs: {"age_days": 572, "days_push": 13, "days_rel": 15, "gap_med": 0, "n_releases_24m": 44}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1178, forks 119 (observed 2026-08-28T04:03:53.050817+00:00)

## What it is
OpenEvals is a library of readymade evaluators for testing LLM applications, including LLM-as-judge evaluators and reference-based/reference-free checks. It is available in both Python and TypeScript and is intended as a starting point for building custom application-specific evals.

## Use cases
- evaluate llm app outputs for correctness
- run llm-as-judge evaluations
- test rag pipeline answer quality
- check if model responses are concise
- grade llm outputs against reference answers
- build custom evaluators for ai agents

## When to choose
- you need off-the-shelf evaluators for LLM app testing
- you want LLM-as-judge scoring with minimal setup
- you use LangChain/LangSmith and want compatible evals

## When to avoid
- you need agent-trajectory-specific evaluation (use agentevals instead)
- you need deterministic unit tests without an LLM judge
- you need a full eval orchestration platform rather than evaluator primitives

## Facets
- artifact type: library
- maturity: active
- function: testing, llm-inference, prompt-engineering, benchmarking
- domain: large-language-models, machine-learning, developer-tools, testing
- platform: python, cross-platform
- tags: llm-evaluation, llm-as-judge, evals, rag-evaluation, langchain, nodejs

## Member repositories
- langchain-ai/openevals (main) score 86

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:53.050817+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 2026-08-30T06:26:06.985589+00:00, confidence not recorded.
  - readme: https://github.com/langchain-ai/openevals (fetched 2026-08-28T04:03:53.050817+00:00, sha dde0264402e1)
- Data as of 2026-08-30T08:39:29.467469+00:00.