Skip to content

Benchmark

LiveBench code_generation

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

Models measured
51
Scores recorded
51
Sources
1
IQ weight
2.19

Score distribution

76.4%
Mean
76.1%
Median
5.6%
Std dev
69%
10th percentile
81.7%
90th percentile
63.4%
Min
63.4%93%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.19 for chat models, measured over 51 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Sonnet 5.5xhigh· LiveBench data93%
2Claude Opus 5.5xhigh· LiveBench data91.5%
3Claude Opus 4.7xhigh· LiveBench data85.9%
4Claude Sonnet 5xhigh· LiveBench data83.1%
5Gemini 3.7 Flashhigh· LiveBench data81.7%
6GPT 5.5xhigh· LiveBench data81.7%
7Muse Spark 1.3xhigh· LiveBench data81.7%
8Claude Opus 4.6auto, high· LiveBench data80.3%
9Claude Sonnet 4.6auto, medium· LiveBench data80.3%
10Gemini 3.5 Flashhigh· LiveBench data80.3%
11GPT-5.2 Codex· LiveBench data80.3%
12Kimi K3· LiveBench data80.3%
13Qwen3.6 Plus· LiveBench data80.3%
14Claude Haiku 5.5xhigh· LiveBench data78.9%
15Claude Opus 4.564k, high· LiveBench data78.9%
16GLM 5.2· LiveBench data78.9%
17GPT-6.1 Solxhigh· LiveBench data78.9%
18Kimi K2.6thinking· LiveBench data78.9%
19Qwen3.7 Max· LiveBench data78.9%
20Gemini 3.6 Flashhigh· LiveBench data77.5%
21GLM 5.3· LiveBench data77.5%
22GLM 5.3 Flash· LiveBench data77.5%
23Grok 4.6· LiveBench data77.5%
24Qwen 3.8 27B· LiveBench data77.5%
25DeepSeek V4 Flash 0731· LiveBench data76.1%
26DeepSeek V4 Pro 0813· LiveBench data76.1%
27Gemini 3.5 Flash Litehigh· LiveBench data76.1%
28GPT 5.2high· LiveBench data76.1%
29Grok 4.7xhigh· LiveBench data76.1%
30Mistral Large 4high· LiveBench data76.1%
31Muse Spark 1.1xhigh· LiveBench data76.1%
32Gemini 3.1 Pro Preview· LiveBench data74.6%
33GPT 5.4xhigh· LiveBench data74.6%
34Grok 4.3· LiveBench data74.6%
35Inkling FP4xhigh· LiveBench data74.6%
36Muse Spark 1.2xhigh· LiveBench data74.6%
37GPT 5.4 Nanoxhigh· LiveBench data74.3%
38DeepSeek V4 Flash· LiveBench data73.2%
39Gemini 3.8 Flashhigh· LiveBench data73.2%
40Kimi K2.7 Code· LiveBench data71.8%
41Nemotron 3 Ultra· LiveBench data71.8%
42Qwen3.6 27B· LiveBench data71.8%
43Qwen3.8 Max· LiveBench data71.8%
44DeepSeek V4 Pro· LiveBench data70.4%
45DeepSeek V4 Flash Vision Exp· LiveBench data69%
46MiniMax M3· LiveBench data69%
47Qwen 3.8 Flash Next· LiveBench data69%
48GPT-5.4 minixhigh· LiveBench data67.7%
49Grok 4.5· LiveBench data67.6%
50Nemotron 3.5 Lightning 30B A3B· LiveBench data67.6%

Showing 1–50 of 51 results.