Skip to content

Benchmark

LiveBench IF

Scores are percentages.

Models measured
42
Scores recorded
46
Sources
1
IQ weight
1.59

Score distribution

69.4%
Mean
69.9%
Median
13.1%
Std dev
54.1%
10th percentile
83.2%
90th percentile
35.6%
Min
35.6%93.3%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.59 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
1GPT 5.1high· Epoch AI Benchmarking Hub data93.3%
2Gemini 2.0 Flash· Epoch AI Benchmarking Hub data85.8%
3o3 Minihigh· Epoch AI Benchmarking Hub data84.4%
4gemini-2.0-pro-exp-02-05· Epoch AI Benchmarking Hub data83.4%
5o3 Minimedium· Epoch AI Benchmarking Hub data83.2%
6Llama 3.3 70B Instruct· Epoch AI Benchmarking Hub data82.7%
7Gemini 2.0 Flash Experimentalthinking· Epoch AI Benchmarking Hub data82.5%
8Gemini 2.0 Flash Experimental· Epoch AI Benchmarking Hub data81.9%
9QwQ 32B· Epoch AI Benchmarking Hub data81.8%
10o1high· Epoch AI Benchmarking Hub data81.6%
11DeepSeek V3· Epoch AI Benchmarking Hub data81.5%
12Claude Sonnet 3.7· Epoch AI Benchmarking Hub data81.3%
13Gemini 2.5 Pro· Epoch AI Benchmarking Hub data80.6%
14DeepSeek R1· Epoch AI Benchmarking Hub data80.5%
15o3 Minilow· Epoch AI Benchmarking Hub data80.1%
16Gemini 2.0 Flash-Lite Preview· Epoch AI Benchmarking Hub data78.3%
17gemini-exp· Epoch AI Benchmarking Hub data77.3%
18Gemini 2.0 Flash Lite· Epoch AI Benchmarking Hub data76.6%
19Perplexity Sonar· Epoch AI Benchmarking Hub data76.2%
20Qwen2.5-Max· Epoch AI Benchmarking Hub data75.4%
21Gemma 3 27B IT· Epoch AI Benchmarking Hub data74.9%
22GPT-4.5 Preview· Epoch AI Benchmarking Hub data72.3%
23DeepSeek R1 Distill Llama 70b· Epoch AI Benchmarking Hub data69.9%
24Grok 2· Epoch AI Benchmarking Hub data69.6%
25Claude 3.5 Sonnet· Epoch AI Benchmarking Hub data69.3%
26GPT 4o· Epoch AI Benchmarking Hub data68.6%
27Learnlm 1.5 Pro Experimental· Epoch AI Benchmarking Hub data68.2%
28Mistral Large2411· Epoch AI Benchmarking Hub data67.9%
29Nova Prov1· Epoch AI Benchmarking Hub data67.1%
30O1 Minimedium· Epoch AI Benchmarking Hub data65.4%
31Claude 3 Opus· Epoch AI Benchmarking Hub data63.9%
32Mistral Small2503· Epoch AI Benchmarking Hub data63.7%
33Claude Haiku 3.5· Epoch AI Benchmarking Hub data61.9%
34Mistral Small· Epoch AI Benchmarking Hub data59.5%
35Qwen 2.5 Coder 32B Instruct· Epoch AI Benchmarking Hub data58.7%
36Phi 4· Epoch AI Benchmarking Hub data58.4%
37Gemma 2 27B IT· Epoch AI Benchmarking Hub data58.1%
38GPT-4o mini· Epoch AI Benchmarking Hub data56.8%
39DeepSeek R1 Distill Qwen 32B· Epoch AI Benchmarking Hub data55.7%
40C4AI Command R 08 2024· Epoch AI Benchmarking Hub data55.6%
41Nova Litev1· Epoch AI Benchmarking Hub data54.1%
42Gemma 2 9B IT· Epoch AI Benchmarking Hub data52.6%
43Nova Microv1· Epoch AI Benchmarking Hub data48%
44Phi-3 Small 8K Instruct· Epoch AI Benchmarking Hub data47.2%
45Phi-3 Mini 4K Instruct· Epoch AI Benchmarking Hub data39.1%
46QwQ 32B Preview· Epoch AI Benchmarking Hub data35.6%

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