Skip to content

Benchmark

LiveCodeBench Codeforces

Coding benchmarks compare models on programming tasks. Scores are percentages.

Models measured
16
Scores recorded
22
Sources
1
IQ weight
2.65

Score distribution

77%
Mean
88.9%
Median
27%
Std dev
17.8%
10th percentile
100%
90th percentile
12.2%
Min
12.2%100%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.65 for chat models, measured over 16 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Gemini 2.5 Flashplatform:codeforces· LiveCodeBench Leaderboard data100%
2Gemini 2.5 Proplatform:codeforces· LiveCodeBench Leaderboard data100%
3o4 Miniplatform:codeforces, high· LiveCodeBench Leaderboard data100%
4o4 Miniplatform:codeforces, low· LiveCodeBench Leaderboard data100%
5o4 Miniplatform:codeforces, medium· LiveCodeBench Leaderboard data100%
6Qwen 3 235B A22Bplatform:codeforces· LiveCodeBench Leaderboard data100%
7DeepSeek R1platform:codeforces· LiveCodeBench Leaderboard data96.7%
8Claude Opus 4platform:codeforces, thinking· LiveCodeBench Leaderboard data88.9%
9Claude Sonnet 4platform:codeforces, thinking· LiveCodeBench Leaderboard data88.9%
10Grok 3 Miniplatform:codeforces, high· LiveCodeBench Leaderboard data88.9%
11o3 Miniplatform:codeforces, high· LiveCodeBench Leaderboard data88.9%
12OpenAI o3platform:codeforces, high· LiveCodeBench Leaderboard data88.9%
13Claude Opus 4platform:codeforces· LiveCodeBench Leaderboard data77.8%
14Claude Sonnet 4platform:codeforces· LiveCodeBench Leaderboard data77.8%
15o3 Miniplatform:codeforces, low· LiveCodeBench Leaderboard data77.8%
16o3 Miniplatform:codeforces, Med· LiveCodeBench Leaderboard data77.8%
17DeepSeek V3platform:codeforces· LiveCodeBench Leaderboard data73.3%
18Claude 3.5 Sonnetplatform:codeforces· LiveCodeBench Leaderboard data53.3%
19GPT-4o miniplatform:codeforces· LiveCodeBench Leaderboard data43.3%
20Claude 3 Haikuplatform:codeforces· LiveCodeBench Leaderboard data17.8%
21GPT 4oplatform:codeforces· LiveCodeBench Leaderboard data17.8%
22GPT-4 Turboplatform:codeforces· LiveCodeBench Leaderboard data12.2%