Skip to content

Benchmark

LiveBench tablejoin

Scores are percentages.

Models measured
51
Scores recorded
51
Sources
1
IQ weight
1.44

Score distribution

47.1%
Mean
46.9%
Median
3.3%
Std dev
42.7%
10th percentile
51%
90th percentile
40.5%
Min
40.5%55.9%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.44 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
1GPT 5.5xhigh· LiveBench data55.9%
2GPT-6.1 Solxhigh· LiveBench data55.3%
3Gemini 3.1 Pro Preview· LiveBench data52.3%
4GPT 5.4xhigh· LiveBench data51.8%
5Qwen 3.8 Flash Next· LiveBench data51%
6Claude Opus 5.5xhigh· LiveBench data51%
7GPT 5.4 Nanoxhigh· LiveBench data50.6%
8Gemini 3.5 Flashhigh· LiveBench data50.5%
9DeepSeek V4 Flash Vision Exp· LiveBench data50.3%
10Qwen3.8 Max· LiveBench data50.1%
11DeepSeek V4 Pro 0813· LiveBench data49.5%
12GPT-5.4 minixhigh· LiveBench data49.4%
13Claude Haiku 5.5xhigh· LiveBench data49.2%
14Mistral Large 4high· LiveBench data49.1%
15GLM 5.3 Flash· LiveBench data48.8%
16Qwen 3.8 27B· LiveBench data48.8%
17DeepSeek V4 Flash 0731· LiveBench data48.5%
18Inkling FP4xhigh· LiveBench data48.5%
19Kimi K3· LiveBench data48.4%
20DeepSeek V4 Pro· LiveBench data48.2%
21Muse Spark 1.3xhigh· LiveBench data48.2%
22Claude Opus 4.6auto, high· LiveBench data48.2%
23Kimi K2.7 Code· LiveBench data47.5%
24Claude Opus 4.7xhigh· LiveBench data47.2%
25Gemini 3.6 Flashhigh· LiveBench data47.1%
26Claude Sonnet 5.5xhigh· LiveBench data46.9%
27Muse Spark 1.2xhigh· LiveBench data46.8%
28Muse Spark 1.1xhigh· LiveBench data46.6%
29Nemotron 3 Ultra· LiveBench data46.6%
30DeepSeek V4 Flash· LiveBench data46.1%
31Kimi K2.6thinking· LiveBench data46.1%
32Claude Opus 4.564k, high· LiveBench data45.9%
33GLM 5.2· LiveBench data45.9%
34MiniMax M3· LiveBench data45.8%
35GPT-5.2 Codex· LiveBench data45.6%
36GPT 5.2high· LiveBench data45.6%
37Qwen3.7 Max· LiveBench data45.5%
38Gemini 3.7 Flashhigh· LiveBench data45.5%
39Gemini 3.8 Flashhigh· LiveBench data45.3%
40Claude Sonnet 4.6auto, medium· LiveBench data45.1%
41Grok 4.7xhigh· LiveBench data44.9%
42Qwen3.6 Plus· LiveBench data44.7%
43Grok 4.5· LiveBench data43.6%
44Grok 4.6· LiveBench data43.4%
45Qwen3.6 27B· LiveBench data42.8%
46Grok Build 0.1· LiveBench data42.7%
47Claude Sonnet 5xhigh· LiveBench data42.4%
48GLM 5.3· LiveBench data42.3%
49Grok 4.3· LiveBench data41.3%
50Gemini 3.5 Flash Litehigh· LiveBench data40.6%

Showing 1–50 of 51 results.