# open-compass/opencompass

OpenCompass is an LLM evaluation platform, supporting a wide range of models (Llama3, Mistral, InternLM2,GPT-4,LLaMa2, Qwen,GLM, Claude, etc) over 100+ datasets.

Repository: https://github.com/open-compass/opencompass
Canonical: https://ross.abutalabs.com/products/opencompass
Homepage: https://opencompass.org.cn/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: evaluation, benchmark, large-language-model, chatgpt, llm, llama2, openai, llama3
Last push: 2026-08-26T08:35:21+00:00

## Health v2 (maintenance only)
Score: 96/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 99, longevity 83
- inputs: {"age_days": 1175, "days_push": 7, "days_rel": 7, "gap_med": 24.0, "n_releases_24m": 17}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 7365, forks 849 (observed 2026-08-28T04:09:58.503778+00:00)

## What it is
OpenCompass is a Python-based evaluation platform and toolkit for benchmarking large language models across 100+ datasets, supporting models like Llama3, Mistral, Qwen, GPT-4, and Claude. It provides configurable evaluation pipelines, leaderboards (CompassRank), and a benchmark hub for comparing LLM quality.

## Use cases
- evaluate llm performance across benchmarks
- compare open-source and proprietary language models
- run standardized llm evaluation suites
- benchmark my fine-tuned model on mmlu and other datasets
- build an llm leaderboard
- test chatgpt-style models on nlp tasks

## When to choose
- you need reproducible, large-scale LLM benchmarking across many datasets
- you want to compare heterogeneous models (OpenAI APIs, HuggingFace checkpoints) under one framework
- you need a maintained, widely-adopted evaluation toolkit with an active community

## When to avoid
- you only need quick ad-hoc prompt testing rather than systematic benchmarking
- you need evaluation of non-text modalities like images or audio
- you want a hosted no-setup evaluation service rather than a self-run framework

## Facets
- artifact type: framework
- maturity: active
- function: benchmarking, machine-learning, llm-inference, nlp, testing
- domain: large-language-models, machine-learning, artificial-intelligence, developer-tools
- platform: python, cli, cross-platform
- tags: llm-evaluation, benchmarking-framework, model-evaluation, openai-api, huggingface-models, natural-language-processing, linux, macos

## Member repositories
- open-compass/opencompass (main) score 96

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:58.503778+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 2026-08-29T17:38:24.622439+00:00, confidence not recorded.
  - readme: https://github.com/open-compass/opencompass (fetched 2026-08-28T04:09:58.503778+00:00, sha f36a1b2cc9b0)
  - homepage: https://opencompass.org.cn/ (fetched 2026-08-29T08:33:52.765539+00:00, sha 1ddf3c7eb5f6)
  - registry_pypi: https://pypi.org/pypi/opencompass/json (fetched 2026-08-29T08:33:52.768606+00:00, sha 2c755df9295c)
- Data as of 2026-08-30T08:39:29.467469+00:00.