---
title: "LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code"
license: CC-BY-4.0
license_url: https://creativecommons.org/licenses/by/4.0/
license_note: 'paper is CC BY 4.0 (verified on arXiv abs page 2026-11-12; icon by-4.0.png, http://creativecommons.org/licenses/by/4.0/); note is our own summary'
source: https://iinventedlight.com/library/notes/2403-07974-livecodebench.md
source_license: 'CC BY 4.0'
attribution: 'Summary by iinventedlight (AI-assisted). Source: Naman Jain, King Han, Alex Gu, and 7 coauthors, LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code (CC BY 4.0).'
author: 'iinventedlight (AI-assisted)'
date: 2026-11-12
status: published
approved_by: "Cole Withington, 2026-11-12 14:44 PT, Librarian chat widget t51s316: \"Yes, publish all 100 second-list arXiv notes\""
checked_on: 2026-11-12
as_of: 2026-11-12
confidence: medium
untested: "Not reproduced. Abs page and HTML key sections read 2026-11-12. Named code URLs checked for HTTP 200; code not run."
category: arxiv
topic_area: 'ai-research: evals'
tags: arxiv, evals, code-generation, contamination, livecodebench
arxiv_id: 2403.07974
alias: 2403-07974-livecodebench
sources:
  - id: S1
    title: "LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code"
    publisher: arXiv
    authors: "Naman Jain and 9 coauthors (listed on the arXiv abs page)"
    url: https://arxiv.org/abs/2403.07974
    version_read: "v2 (6 Jun 2024)"
    accessed: 2026-11-12
    license: 'CC BY 4.0'
---
# LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code

## Summary
LiveCodeBench is a contamination-aware benchmark for code-related language models. It continuously collects new problems over time from contests across LeetCode, AtCoder, and CodeForces. The benchmark covers more than generation: self-repair, code execution, and test output prediction. The v2 paper reports four hundred problems published between May 2023 and May 2024, and evaluations of 18 base and 34 instruction-tuned models, plus findings on contamination and possible overfitting on older sets such as HumanEval. The comments give the website https://livecodebench.github.io/. [S1]

## Claims
- id: C1
  claim: "LiveCodeBench continuously collects new contest problems from LeetCode, AtCoder, and CodeForces."
  source: S1
  evidence: "continuously collects new problems over time from contests"
- id: C2
  claim: "The v2 release hosts four hundred problems published between May 2023 and May 2024."
  source: S1
  evidence: "four hundred high-quality coding problems that were published between May 2023 and May 2024"
- id: C3
  claim: "The authors evaluated 18 base models and 34 instruction-tuned models, and point to a public website."
  source: S1
  evidence: "https://livecodebench.github.io/"

## Verification
- Code available: Paper comments and abs page: https://livecodebench.github.io/, which returned HTTP 200 on 2026-11-12. Not run.
- Reproduced: not reproduced (2026-11-12). Code not run.
- Related work: Live library notes on earlier code evals include [2107-03374-humaneval](https://iinventedlight.com/library/notes/2107-03374-humaneval/) and [2310-06770-swe-bench](https://iinventedlight.com/library/notes/2310-06770-swe-bench/). See also [2402-01030-codeact](https://iinventedlight.com/library/notes/2402-01030-codeact/).
- Limits: Coverage is contest-style programming, not full software-engineering workflows. Problem counts and model lists are as of the v2 paper (6 Jun 2024) and will change as the live set grows.
- License: CC BY 4.0 on the abs page (2026-11-12; icon by-4.0.png; http://creativecommons.org/licenses/by/4.0/). Submitted 12 Mar 2024 (v1); last revised v2 (6 Jun 2024). Sources: [S1] https://arxiv.org/abs/2403.07974, v2 (6 Jun 2024). checked_on: 2026-11-12.

## When to apply
Use when asking how to evaluate code models on problems that post-date a training cutoff, or how LiveCodeBench differs from HumanEval.

## Facts
- id: F1
  claim: "LiveCodeBench collects new coding problems over time from LeetCode, AtCoder, and CodeForces contests."
  confidence: medium
  date: 2026-11-12
  sources: [S1]
  license: CC BY 4.0
- id: F2
  claim: "The v2 paper reports four hundred problems published between May 2023 and May 2024, and evaluations of 18 base and 34 instruction-tuned models."
  confidence: medium
  date: 2026-11-12
  sources: [S1]
  license: CC BY 4.0
- id: F3
  claim: "The project website is https://livecodebench.github.io/."
  confidence: medium
  date: 2026-11-12
  sources: [S1]
  license: CC BY 4.0

## Sources
- [S1] arXiv 2403.07974, https://arxiv.org/abs/2403.07974. License: CC BY 4.0. v2 (6 Jun 2024). Accessed 2026-11-12.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Summary by iinventedlight (AI-assisted). Paper is CC BY 4.0.

---

Agents: [/agents.md](/agents.md) | [/.well-known/agent-library.json](/.well-known/agent-library.json).
