Skip to content

Benchmark

LiveBench logic_with_navigation

Scores are percentages.

Models measured
51
Scores recorded
51
Sources
1
IQ weight
2.55

Score distribution

70.7%
Mean
70%
Median
10.2%
Std dev
64%
10th percentile
82%
90th percentile
18%
Min
18%88%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.55 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
1Muse Spark 1.2xhigh· LiveBench data88%
2Claude Sonnet 5xhigh· LiveBench data86%
3Qwen3.7 Max· LiveBench data84%
4Gemini 3.8 Flashhigh· LiveBench data82%
5GPT-6.1 Solxhigh· LiveBench data82%
6Grok 4.6· LiveBench data82%
7Claude Opus 4.6auto, high· LiveBench data80%
8Claude Opus 5.5xhigh· LiveBench data80%
9Claude Sonnet 5.5xhigh· LiveBench data80%
10GLM 5.2· LiveBench data80%
11Kimi K3· LiveBench data80%
12Muse Spark 1.1xhigh· LiveBench data78%
13Gemini 3.6 Flashhigh· LiveBench data76%
14GPT 5.5xhigh· LiveBench data76%
15Muse Spark 1.3xhigh· LiveBench data76%
16Gemini 3.5 Flashhigh· LiveBench data74%
17Gemini 3.7 Flashhigh· LiveBench data74%
18GPT 5.2high· LiveBench data74%
19Kimi K2.7 Code· LiveBench data74%
20Qwen3.8 Max· LiveBench data74%
21Gemini 3.1 Pro Preview· LiveBench data72%
22Grok 4.5· LiveBench data72%
23Claude Opus 4.7xhigh· LiveBench data70%
24Claude Sonnet 4.6auto, medium· LiveBench data70%
25DeepSeek V4 Flash 0731· LiveBench data70%
26GLM 5.3· LiveBench data70%
27Grok Build 0.1· LiveBench data70%
28Kimi K2.6thinking· LiveBench data70%
29Qwen 3.8 Flash Next· LiveBench data70%
30Qwen3.6 Plus· LiveBench data70%
31Claude Opus 4.564k, high· LiveBench data68%
32DeepSeek V4 Flash Vision Exp· LiveBench data68%
33Gemini 3.5 Flash Litehigh· LiveBench data68%
34GPT 5.4xhigh· LiveBench data68%
35GPT-5.2 Codex· LiveBench data68%
36Mistral Large 4high· LiveBench data68%
37Nemotron 3 Ultra· LiveBench data68%
38Qwen 3.8 27B· LiveBench data68%
39DeepSeek V4 Pro 0813· LiveBench data66%
40Grok 4.3· LiveBench data66%
41Grok 4.7xhigh· LiveBench data66%
42Inkling FP4xhigh· LiveBench data66%
43MiniMax M3· LiveBench data66%
44DeepSeek V4 Flash· LiveBench data64%
45DeepSeek V4 Pro· LiveBench data64%
46GPT 5.4 Nanoxhigh· LiveBench data64%
47GLM 5.3 Flash· LiveBench data62%
48Qwen3.6 27B· LiveBench data62%
49GPT-5.4 minixhigh· LiveBench data60%
50Claude Haiku 5.5xhigh· LiveBench data56%

Showing 1–50 of 51 results.