Skip to content

Benchmark

LiveBench code_completion

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

Models measured
51
Scores recorded
51
Sources
1
IQ weight
2.07

Score distribution

74.9%
Mean
76.1%
Median
9%
Std dev
67.4%
10th percentile
82.6%
90th percentile
23.9%
Min
23.9%87%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.07 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 Opus 5.5xhigh· LiveBench data87%
2GPT-5.2 Codex· LiveBench data87%
3Claude Sonnet 5.5xhigh· LiveBench data84.8%
4GPT 5.5xhigh· LiveBench data82.6%
5GPT-6.1 Solxhigh· LiveBench data82.6%
6Kimi K3· LiveBench data82.6%
7Claude Opus 4.564k, high· LiveBench data80.4%
8GLM 5.2· LiveBench data80.4%
9GLM 5.3· LiveBench data80.4%
10GLM 5.3 Flash· LiveBench data80.4%
11GPT 5.4xhigh· LiveBench data80.4%
12Muse Spark 1.2xhigh· LiveBench data80.4%
13Muse Spark 1.3xhigh· LiveBench data80.4%
14Claude Opus 4.7xhigh· LiveBench data78.3%
15Claude Sonnet 4.6auto, medium· LiveBench data78.3%
16Claude Sonnet 5xhigh· LiveBench data78.3%
17DeepSeek V4 Pro 0813· LiveBench data78.3%
18Gemini 3.1 Pro Preview· LiveBench data78.3%
19Gemini 3.6 Flashhigh· LiveBench data78.3%
20Grok 4.7xhigh· LiveBench data78.3%
21Kimi K2.6thinking· LiveBench data78.3%
22Mistral Large 4high· LiveBench data78.3%
23Muse Spark 1.1xhigh· LiveBench data78.3%
24Claude Opus 4.6auto, high· LiveBench data76.1%
25Gemini 3.5 Flashhigh· LiveBench data76.1%
26Gemini 3.5 Flash Litehigh· LiveBench data76.1%
27Gemini 3.7 Flashhigh· LiveBench data76.1%
28GPT 5.2high· LiveBench data76.1%
29Grok 4.6· LiveBench data76.1%
30Kimi K2.7 Code· LiveBench data76.1%
31Qwen 3.8 Flash Next· LiveBench data76.1%
32Qwen3.6 Plus· LiveBench data76.1%
33GPT-5.4 minixhigh· LiveBench data75.6%
34Claude Haiku 5.5xhigh· LiveBench data73.9%
35DeepSeek V4 Flash 0731· LiveBench data73.9%
36Qwen 3.8 27B· LiveBench data73.9%
37Qwen3.8 Max· LiveBench data73.9%
38Gemini 3.8 Flashhigh· LiveBench data71.7%
39Qwen3.6 27B· LiveBench data71.7%
40DeepSeek V4 Pro· LiveBench data69.6%
41Grok 4.5· LiveBench data69.6%
42Nemotron 3 Ultra· LiveBench data69.6%
43Qwen3.7 Max· LiveBench data69.6%
44DeepSeek V4 Flash Vision Exp· LiveBench data67.4%
45GPT 5.4 Nanoxhigh· LiveBench data67.4%
46Grok Build 0.1· LiveBench data67.4%
47Inkling FP4xhigh· LiveBench data67.4%
48MiniMax M3· LiveBench data67.4%
49DeepSeek V4 Flash· LiveBench data65.2%
50Grok 4.3· LiveBench data65.2%

Showing 1–50 of 51 results.