Skip to content

Benchmark

LiveBench Data Analysis

Scores are percentages.

Models measured
42
Scores recorded
46
Sources
1
IQ weight
1.95

Score distribution

55.1%
Mean
55.9%
Median
12.4%
Std dev
36.4%
10th percentile
69.8%
90th percentile
30.3%
Min
30.3%79.9%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.95 for chat models, measured over 41 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 Pro· Epoch AI Benchmarking Hub data79.9%
2Claude Sonnet 3.7· Epoch AI Benchmarking Hub data74%
3GPT 5.1high· Epoch AI Benchmarking Hub data72.1%
4o3 Minihigh· Epoch AI Benchmarking Hub data70.6%
5DeepSeek R1· Epoch AI Benchmarking Hub data69.8%
6Gemini 2.0 Flash Experimentalthinking· Epoch AI Benchmarking Hub data69.4%
7gemini-2.0-pro-exp-02-05· Epoch AI Benchmarking Hub data68%
8Qwen2.5-Max· Epoch AI Benchmarking Hub data67.9%
9Gemini 2.0 Flash· Epoch AI Benchmarking Hub data67.6%
10o3 Minimedium· Epoch AI Benchmarking Hub data66.6%
11o1high· Epoch AI Benchmarking Hub data65.5%
12Gemini 2.0 Flash Lite· Epoch AI Benchmarking Hub data65.5%
13QwQ 32B· Epoch AI Benchmarking Hub data65%
14GPT-4.5 Preview· Epoch AI Benchmarking Hub data64.3%
15gemini-exp· Epoch AI Benchmarking Hub data63.2%
16o3 Minilow· Epoch AI Benchmarking Hub data62%
17Gemini 2.0 Flash Experimental· Epoch AI Benchmarking Hub data61.7%
18GPT 4o· Epoch AI Benchmarking Hub data60.9%
19DeepSeek V3· Epoch AI Benchmarking Hub data60.4%
20O1 Minimedium· Epoch AI Benchmarking Hub data57.9%
21Claude 3 Opus· Epoch AI Benchmarking Hub data57.9%
22Gemini 2.0 Flash-Lite Preview· Epoch AI Benchmarking Hub data57.5%
23DeepSeek R1 Distill Llama 70b· Epoch AI Benchmarking Hub data55.9%
24Claude 3.5 Sonnet· Epoch AI Benchmarking Hub data55%
25Learnlm 1.5 Pro Experimental· Epoch AI Benchmarking Hub data55%
26Grok 2· Epoch AI Benchmarking Hub data54.5%
27Mistral Small· Epoch AI Benchmarking Hub data53.7%
28Gemma 3 27B IT· Epoch AI Benchmarking Hub data51.5%
29Mistral Small2503· Epoch AI Benchmarking Hub data50.5%
30Mistral Large2411· Epoch AI Benchmarking Hub data50.2%
31GPT-4o mini· Epoch AI Benchmarking Hub data50%
32Qwen 2.5 Coder 32B Instruct· Epoch AI Benchmarking Hub data49.9%
33Llama 3.3 70B Instruct· Epoch AI Benchmarking Hub data49.5%
34Claude Haiku 3.5· Epoch AI Benchmarking Hub data48.5%
35Nova Prov1· Epoch AI Benchmarking Hub data48.3%
36Gemma 2 27B IT· Epoch AI Benchmarking Hub data47.9%
37DeepSeek R1 Distill Qwen 32B· Epoch AI Benchmarking Hub data45.4%
38Phi 4· Epoch AI Benchmarking Hub data45.2%
39Perplexity Sonar· Epoch AI Benchmarking Hub data37.9%
40Nova Litev1· Epoch AI Benchmarking Hub data37.2%
41Gemma 2 9B IT· Epoch AI Benchmarking Hub data36.4%
42Phi-3 Mini 4K Instruct· Epoch AI Benchmarking Hub data34.7%
43Nova Microv1· Epoch AI Benchmarking Hub data34%
44C4AI Command R 08 2024· Epoch AI Benchmarking Hub data33.3%
45QwQ 32B Preview· Epoch AI Benchmarking Hub data31.6%
46Phi-3 Small 8K Instruct· Epoch AI Benchmarking Hub data30.3%

Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence)