Skip to content

Benchmark

NYT Connections – Extended Version (Lech Mazur)

Scores are percentages.

Models measured
106
Scores recorded
138
Sources
1
IQ weight
2.44

Score distribution

63.5%
Mean
75.1%
Median
30.5%
Std dev
14.5%
10th percentile
94.4%
90th percentile
4.1%
Min
4.1%99.3%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.44 for chat models, measured over 106 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 4 Carbonhigh reasoning· Lech Mazur suite data99.3%
2Gemini 4 Neonhigh reasoning· Lech Mazur suite data99.1%
3Gemini 4 Argonhigh reasoning· Lech Mazur suite data98.6%
4GPT-6 Astramax reasoning· Lech Mazur suite data98.4%
5GPT-6.2 Astrahigh reasoning· Lech Mazur suite data98.2%
6GPT-6 Astraxhigh reasoning· Lech Mazur suite data98.1%
7GPT-6.1 Solmax reasoning· Lech Mazur suite data97.8%
8GPT-6 Astrahigh reasoning· Lech Mazur suite data97.7%
9Gemini 3.1 Pro Preview· Lech Mazur suite data97.4%
10Gemini 3.8 Flashhigh reasoning· Lech Mazur suite data97.4%
11Claude Fable 5.2high reasoning· Lech Mazur suite data97.3%
12GPT 5.5xhigh reasoning· Lech Mazur suite data96.2%
13GPT-6.1 Solhigh reasoning· Lech Mazur suite data95.5%
14GPT 5.5high reasoning· Lech Mazur suite data95.2%
15Gemini 3 Pro Preview· Lech Mazur suite data94.4%
16Claude Opus 5xhigh reasoning· Lech Mazur suite data94.3%
17GPT-6.2 Solhigh reasoning· Lech Mazur suite data94.1%
18Gemini 3.7 Flash· Lech Mazur suite data94%
19GPT-5.6 Solxhigh reasoning· Lech Mazur suite data93.8%
20Kimi K3· Lech Mazur suite data93.6%
21Kimi K3.1 Prohigh reasoning· Lech Mazur suite data93.4%
22Muse Wonder 1.0high reasoning· Lech Mazur suite data93.4%
23DeepSeek V4.1 Flash 1017high reasoning· Lech Mazur suite data92.8%
24Claude Fable 5high reasoning· Lech Mazur suite data92.7%
25Gemini 3.5 Flash· Lech Mazur suite data92.6%
26GPT 5.5medium reasoning· Lech Mazur suite data92.4%
27Claude Opus 5high reasoning· Lech Mazur suite data92.2%
28Claude Opus 4.6high reasoning· Lech Mazur suite data92.1%
29DeepSeek V4 Prohigh reasoning· Lech Mazur suite data91.3%
30GPT 5.4xhigh reasoning· Lech Mazur suite data91.3%
31Claude Opus 4.8xhigh reasoning· Lech Mazur suite data91.1%
32GPT-5.6 Solhigh reasoning· Lech Mazur suite data91%
33Qwen4 Maxhigh· Lech Mazur suite data90.8%
34GPT 5.4high reasoning· Lech Mazur suite data90.6%
35GPT-6 Solhigh reasoning· Lech Mazur suite data90.1%
36Claude Fable 5.1high reasoning· Lech Mazur suite data90%
37DeepSeek V4 Flash· Lech Mazur suite data89.6%
38DeepSeek V4.1 Flashhigh reasoning· Lech Mazur suite data89.6%
39Gemini 3.6 Flash· Lech Mazur suite data89%
40Claude Opus 5.5high reasoning· Lech Mazur suite data88.5%
41Claude Opus 4.8high reasoning· Lech Mazur suite data88.3%
42Qwen3.8 Max· Lech Mazur suite data88.3%
43Claude Opus 5.5medium reasoning· Lech Mazur suite data87.8%
44GPT 5.4medium reasoning· Lech Mazur suite data87.8%
45Grok 4.1 Fastreasoning· Lech Mazur suite data87.4%
46Kimi K2.6· Lech Mazur suite data87.2%
47GPT-6.1 Terrahigh reasoning· Lech Mazur suite data86.8%
48Grok 4.20Reasoning· Lech Mazur suite data85.4%
49Muse Spark 1.3high reasoning· Lech Mazur suite data85.1%
50Qwen3.7 Max· Lech Mazur suite data85.1%

Showing 1–50 of 138 results.