# LiveBench/LiveBench

LiveBench: A Challenging, Contamination-Free LLM Benchmark

Repository: https://github.com/LiveBench/LiveBench
Canonical: https://ross.abutalabs.com/products/livebench
Language: Python
License: NOASSERTION
License Family: other
Last push: 2026-08-26T15:00:50+00:00

## Health v2 (maintenance only)
Score: 68/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 35, longevity 58
- inputs: {"age_days": 812, "days_push": 7, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1294, forks 119 (observed 2026-08-28T04:04:16.349389+00:00)

## What it is
LiveBench is a contamination-free benchmark for large language models that releases new questions monthly, drawn from recent datasets, papers, news, and IMDb synopses. It includes 18 diverse tasks across 6 categories with objective, verifiable ground-truth answers that are scored automatically without an LLM judge.

## Use cases
- benchmark llm performance without test set contamination
- evaluate my model on a public leaderboard
- score llm answers objectively without an llm judge
- run monthly fresh benchmark questions for chat models
- compare api models on coding math and instruction following tasks
- evaluate agentic coding questions in docker

## When to choose
- you need contamination-resistant evaluation of LLMs
- you want automatic objective scoring without judge models
- you want to submit a model to a public leaderboard

## When to avoid
- you need local model inference support, which is unmaintained
- you need a static benchmark with a fixed question set
- you need subjective or open-ended evaluation criteria

## Facets
- artifact type: dataset
- maturity: active
- function: benchmarking, llm-inference, testing
- domain: large-language-models, machine-learning, artificial-intelligence
- platform: python, cli
- tags: llm-benchmark, contamination-free, evaluation, leaderboard, model-evaluation, docker

## Member repositories
- LiveBench/LiveBench (main) score 68

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:16.349389+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 2026-08-30T04:54:03.277043+00:00, confidence not recorded.
  - readme: https://github.com/LiveBench/LiveBench (fetched 2026-08-28T04:04:16.349389+00:00, sha eb1d2aa73347)
- Data as of 2026-08-30T08:39:29.467469+00:00.