Skip to content

Benchmark

Mirrorcode

Scores are percentages.

Models measured
39
Scores recorded
46
Sources
1
IQ weight
1.81

Score distribution

48.5%
Mean
46.7%
Median
23.6%
Std dev
21.1%
10th percentile
78.9%
90th percentile
4.4%
Min
4.4%97.8%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.81 for chat models, measured over 39 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 Sonnet 5.5max· Epoch AI Benchmarking Hub data97.8%
2Claude Opus 5.5max· Epoch AI Benchmarking Hub data95.6%
3Claude Fable 5.1max· Epoch AI Benchmarking Hub data93.3%
4Claude Opus 5max· Epoch AI Benchmarking Hub data87.8%
5Claude Opus 5.5high· Epoch AI Benchmarking Hub data78.9%
6Claude Haiku 5.5max· Epoch AI Benchmarking Hub data78.9%
7GPT-6.1 Terramax· Epoch AI Benchmarking Hub data76.7%
8Claude Fable 5.1high· Epoch AI Benchmarking Hub data73.3%
9Claude Fable 5max· Epoch AI Benchmarking Hub data70%
10GPT-6.2 Astramax· Epoch AI Benchmarking Hub data67.8%
11Claude Opus 4.8max· Epoch AI Benchmarking Hub data65.6%
12DeepSeek V4.1 Flashmax· Epoch AI Benchmarking Hub data64.4%
13Claude Fable 5high· Epoch AI Benchmarking Hub data63.9%
14Claude Sonnet 5max· Epoch AI Benchmarking Hub data63.3%
15GPT-6.1 Solmax· Epoch AI Benchmarking Hub data61.1%
16GPT-6 Astramax· Epoch AI Benchmarking Hub data60%
17Gemini 4 Argonhigh· Epoch AI Benchmarking Hub data58.9%
18Grok 4.7xhigh· Epoch AI Benchmarking Hub data54.4%
19Kimi K3max· Epoch AI Benchmarking Hub data54.4%
20Muse Spark 1.3max· Epoch AI Benchmarking Hub data54.4%
21GPT-6 Solmax· Epoch AI Benchmarking Hub data48.9%
22Qwen3.8 Maxxhigh· Epoch AI Benchmarking Hub data47.8%
23GPT-6 Astrahigh· Epoch AI Benchmarking Hub data46.7%
24Muse Wonder 1.0max· Epoch AI Benchmarking Hub data45.6%
25Claude Opus 4.7max· Epoch AI Benchmarking Hub data43.3%
26Gemini 3.7 Flashhigh· Epoch AI Benchmarking Hub data41.1%
27GPT-5.6 Terramax· Epoch AI Benchmarking Hub data41.1%
28Grok 4.6xhigh· Epoch AI Benchmarking Hub data41.1%
29Mistral Large 4high· Epoch AI Benchmarking Hub data40%
30Gemini 3.8 Flashhigh· Epoch AI Benchmarking Hub data38.9%
31GLM 5.2max· Epoch AI Benchmarking Hub data36.7%
32Claude Opus 4.6max· Epoch AI Benchmarking Hub data33.3%
33DeepSeek V4 Pro 0813max· Epoch AI Benchmarking Hub data33.3%
34GPT 5.2xhigh· Epoch AI Benchmarking Hub data32.2%
35Grok 4.5high· Epoch AI Benchmarking Hub data32.2%
36Claude Opus 4.7high· Epoch AI Benchmarking Hub data31.1%
37GPT-6 Lunamax· Epoch AI Benchmarking Hub data26.7%
38Gemini 3.6 Flashhigh· Epoch AI Benchmarking Hub data23.3%
39GPT 5.4xhigh· Epoch AI Benchmarking Hub data23.3%
40GPT 5.5xhigh· Epoch AI Benchmarking Hub data23.3%
41GPT-5.6 Lunamax· Epoch AI Benchmarking Hub data21.1%
42GPT-5.6 Solhigh· Epoch AI Benchmarking Hub data20%
43GPT 5.4high· Epoch AI Benchmarking Hub data15.6%
44GPT 5.5high· Epoch AI Benchmarking Hub data10%
45Gemini 3.1 Pro Previewhigh· Epoch AI Benchmarking Hub data8.9%
46GPT 5high· Epoch AI Benchmarking Hub data4.4%

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