evals.report
BenchmarksLabsCompareRun guidesIn the wild
BenchmarksCoding

LiveCodeBench

A holistic, contamination-free benchmark that continuously collects new competitive-programming problems from LeetCode, AtCoder, and Codeforces (released after model training cutoffs) and measures code-generation correctness via Pass@1.

CodingPass@1Higher is better
ScoresAboutRun this benchmark

What is LiveCodeBench?

A holistic, contamination-free benchmark that continuously collects new competitive-programming problems from LeetCode, AtCoder, and Codeforces (released after model training cutoffs) and measures code-generation correctness via Pass@1. evals.report tracks reported LiveCodeBench scores with the model, source, status, date, and run caveats attached — official leaderboard scores, vendor-reported launches, and clearly labeled community runs.

Top reported LiveCodeBench score: Fugu Ultra — 93.2% (Pass@1).

ModelLabScore↓Source modelStatusDate
Fugu UltraSakana AI93.2%Fugu UltraVerifiedJun 15, 2026Details
FuguSakana AI92.9%FuguVerifiedJun 15, 2026Details
Gemini 3 ProGoogle DeepMind91.7%—UnverifiedNov 18, 2025Details
GPT-5.2OpenAI89.4%—UnverifiedDec 11, 2025Details
GPT-OSS-120BOpenOpenAI87.8%—UnverifiedAug 5, 2025Details
GPT-5.1OpenAI86.8%—UnverifiedNov 12, 2025Details
o4-miniOpenAI85.9%—UnverifiedApr 16, 2025Details
Kimi K2 InstructOpenMoonshot AI85.3%—UnverifiedJul 11, 2025Details
GPT-5OpenAI84.6%—UnverifiedAug 7, 2025Details
GPT-5 miniOpenAI83.8%—UnverifiedAug 7, 2025Details
Grok 4.1 fast reasoningxAI82.2%—UnverifiedNov 19, 2025Details
Grok 4xAI81.9%—UnverifiedJul 9, 2025Details
MiniMax M2.1OpenMiniMax81.0%—UnverifiedDec 23, 2025Details
o3OpenAI80.8%—UnverifiedApr 16, 2025Details
Gemini 2.5 ProGoogle DeepMind80.1%—UnverifiedMar 25, 2025Details
Gemini 3 FlashGoogle DeepMind79.7%—UnverifiedDec 17, 2025Details
Qwen3 MaxAlibaba / Qwen76.7%—UnverifiedSep 5, 2025Details
Claude Opus 4.5Anthropic73.8%—UnverifiedNov 24, 2025Details
DeepSeek R1OpenDeepSeek61.7%—UnverifiedJan 20, 2025Details
DeepSeek V3.2OpenDeepSeek59.3%—UnverifiedDec 1, 2025Details
Claude Sonnet 4.5Anthropic59.0%—UnverifiedSep 29, 2025Details
Qwen 3 Coder 480BOpenAlibaba / Qwen58.5%—UnverifiedJul 22, 2025Details
DeepSeek V3.1OpenDeepSeek57.7%—UnverifiedAug 21, 2025Details
GLM-4.6OpenZ.ai56.1%—UnverifiedSep 30, 2025Details
Claude Opus 4Anthropic54.2%—UnverifiedMay 22, 2025Details
Qwen3 235B A22B Instruct 2507OpenAlibaba / Qwen52.4%—UnverifiedJul 21, 2025Details
Claude Haiku 4.5Anthropic51.1%—UnverifiedOct 15, 2025Details
Gemini 2.5 FlashGoogle DeepMind49.5%—UnverifiedApr 17, 2025Details
GPT-4.1OpenAI45.7%—UnverifiedApr 14, 2025Details
Claude Sonnet 4Anthropic44.9%—UnverifiedMay 22, 2025Details
DeepSeek V3 0324OpenDeepSeek40.5%—UnverifiedMar 24, 2025Details
Llama 4 MaverickOpenMeta39.7%—UnverifiedApr 5, 2025Details
Claude 3.7 SonnetAnthropic39.4%—UnverifiedFeb 24, 2025Details
Claude 3.5 SonnetAnthropic38.1%—UnverifiedJun 20, 2024Details
Amazon Nova 2 LiteAmazon34.6%—UnverifiedDec 2, 2025Details
Gemini 2.0 FlashGoogle DeepMind33.4%—UnverifiedDec 11, 2024Details
Llama 3.1 405BOpenMeta30.5%—UnverifiedJul 23, 2024Details
Llama 4 ScoutOpenMeta29.9%—UnverifiedApr 5, 2025Details
Jamba 1.7 LargeOpenAI21 Labs18.1%—UnverifiedJul 3, 2025Details
Mistral LargeMistral AI17.8%—UnverifiedFeb 26, 2024Details

Each row reports the model’s Pass@1 on LiveCodeBench. Click a row for the full run context.