LiveCodeBench
LiveCodeBench is a holistic and contamination-free evaluation benchmark for large language models for code. It continuously collects new problems from programming contests (LeetCode, AtCoder, CodeForces) and evaluates four different scenarios: code generation, self-repair, code execution, and test output prediction. Problems are annotated with release dates to enable evaluation on unseen problems released after a model's training cutoff.
DeepSeek-V4-Pro-Max from DeepSeek currently leads the LiveCodeBench leaderboard with a score of 0.935 across 74 evaluated AI models.
What LiveCodeBench measures
LiveCodeBench is a text benchmark that evaluates large language models on reasoning, general, and code tasks. LLM Stats tracks 74 models (1 with independently verified scores) on this benchmark, with a maximum possible score of 1. Current average across reported models is 0.5, with the leader reaching 0.9.
Compare leaders on the best AI for reasoning, best AI for general and best AI for code leaderboards.
Publication
- Paper
- LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
- Authors
- Naman Jain, King Han, Alex Gu, Wen-Ding Li, and 6 others
- Published
- arXiv
- 2403.07974
Abstract
Large Language Models (LLMs) applied to code-related applications have emerged as a prominent field, attracting significant interest from both academia and industry. However, as new and improved LLMs are developed, existing evaluation benchmarks (e.g., HumanEval, MBPP) are no longer sufficient for assessing their capabilities. In this work, we propose LiveCodeBench, a comprehensive and contamination-free evaluation of LLMs for code, which continuously collects new problems over time from contests across three competition platforms, namely LeetCode, AtCoder, and CodeForces. Notably, our benchmark also focuses on a broader range of code related capabilities, such as self-repair, code execution, and test output prediction, beyond just code generation. Currently, LiveCodeBench hosts four hundred high-quality coding problems that were published between May 2023 and May 2024. We have evaluated 18 base LLMs and 34 instruction-tuned LLMs on LiveCodeBench. We present empirical findings on contamination, holistic performance comparisons, potential overfitting in existing benchmarks as well as individual model comparisons. We will release all prompts and model completions for further community analysis, along with a general toolkit for adding new scenarios and model
DeepSeek-V4-Pro-Max leads with 93.5%, followed by DeepSeek-V4-Flash-Max at 91.6% and DeepSeek-V3.2 at 83.3%.
Progress Over Time
Interactive timeline showing model performance evolution on LiveCodeBench
LiveCodeBench Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 | DeepSeek | 1.6T | 1.0M | $1.74 / $3.48 | ||
| 2 | DeepSeek | 284B | 1.0M | $0.14 / $0.28 | ||
| 3 | DeepSeek | 685B | — | — | ||
| 3 | DeepSeek | 685B | — | — | ||
| 5 |  MiniMax | 230B | 1.0M | $0.30 / $1.20 | ||
| 6 |  Meituan | 560B | — | — | ||
| 7 | 120B | — | — | |||
| 8 |  xAI | — | — | — | ||
| 9 | xAI | — | 2.0M | $0.20 / $0.50 | ||
| 10 | xAI | — | — | — | ||
| 10 | xAI | — | 128K | $3.00 / $15.00 | ||
| 10 | Meituan | 560B | — | — | ||
| 13 | xAI | — | — | — | ||
| 14 | MiniMax | 230B | 1.0M | $0.30 / $1.20 | ||
| 15 |  Amazon | — | — | — | ||
| 16 | DeepSeek | 685B | — | — | ||
| 17 | DeepSeek | 671B | 131K | $0.55 / $2.19 | ||
| 18 |  Zhipu AI | 355B | — | — | ||
| 19 |  NVIDIA | 9B | — | — | ||
| 20 | Amazon | — | 1.0M | $0.30 / $2.50 | ||
| 21 | Zhipu AI | 106B | — | — | ||
| 21 |  Alibaba Cloud / Qwen Team | 235B | — | — | ||
| 23 | — | — | — | |||
| 24 |  Inception | — | 128K | $0.25 / $0.75 | ||
| 25 | 253B | — | — | |||
| 26 | Alibaba Cloud / Qwen Team | 33B | 128K | $0.10 / $0.44 | ||
| 27 | MiniMax | 456B | — | — | ||
| 28 |  Mistral AI | 14B | — | — | ||
| 29 | Mistral AI | 119B | 256K | $0.15 / $0.60 | ||
| 30 | Alibaba Cloud / Qwen Team | 33B | — | — | ||
| 31 | Alibaba Cloud / Qwen Team | 31B | 128K | $0.10 / $0.30 | ||
| 32 | MiniMax | 456B | — | — | ||
| 33 | Mistral AI | 8B | — | — | ||
| 34 | DeepSeek | 71B | — | — | ||
| 35 | DeepSeek | 33B | — | — | ||
| 36 | DeepSeek | 671B | — | — | ||
| 37 | Alibaba Cloud / Qwen Team | 73B | — | — | ||
| 38 | Mistral AI | 3B | — | — | ||
| 39 |  Microsoft | 14B | — | — | ||
| 40 |  Moonshot AI | 1.0T | — | — | ||
| 41 | Microsoft | 14B | — | — | ||
| 41 | DeepSeek | 15B | — | — | ||
| 43 | Mistral AI | 24B | — | — | ||
| 44 | Mistral AI | 24B | — | — | ||
| 45 | Alibaba Cloud / Qwen Team | 33B | — | — | ||
| 45 | DeepSeek | 671B | — | — | ||
| 47 | DeepSeek | 671B | 164K | $0.28 / $1.14 | ||
| 48 | Meituan | 560B | — | — | ||
| 49 |  Meta | 400B | — | — | ||
| 50 | DeepSeek | 8B | — | — |
FAQ
Common questions about LiveCodeBench.
More evaluations to explore
Related benchmarks in the same category
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks. Features over 12,000 curated questions across 14 domains and causes a 16-33% accuracy drop compared to original MMLU.
All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.
Massive Multitask Language Understanding benchmark testing knowledge across 57 diverse subjects including STEM, humanities, social sciences, and professional domains
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions