Skip to content

Benchmark

NYT Connections – Older models (Lech Mazur)

Scores are percentages.

Models measured
77
Scores recorded
93
Sources
1
IQ weight
3.37

Score distribution

27.2%
Mean
13.9%
Median
26%
Std dev
4.2%
10th percentile
70.5%
90th percentile
0.7%
Min
0.7%89.5%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 3.37 for chat models, measured over 77 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1GLM 5.2max reasoning· Lech Mazur suite data89.5%
2Qwen3.8 Max· Lech Mazur suite data88.8%
3Grok 4 Fastreasoning· Lech Mazur suite data88%
4Inkling FP4high reasoning· Lech Mazur suite data88%
5Grok 4· Lech Mazur suite data87.8%
6MiMo V2.6 Prothinking· Lech Mazur suite data84%
7O3 Promedium reasoning· Lech Mazur suite data81.7%
8GPT-5 Pro· Lech Mazur suite data77%
9O1 Promedium reasoning· Lech Mazur suite data74.6%
10OpenAI o3high reasoning· Lech Mazur suite data70.5%
11GPT 5high reasoning· Lech Mazur suite data67.8%
12o4 Minihigh reasoning· Lech Mazur suite data64%
13OpenAI o3medium reasoning· Lech Mazur suite data63%
14GPT 5medium reasoning· Lech Mazur suite data61.8%
15o1medium reasoning· Lech Mazur suite data60.1%
16GPT 5.1high reasoning· Lech Mazur suite data59.7%
17o4 Minimedium reasoning· Lech Mazur suite data58.4%
18GPT-5 minimedium reasoning· Lech Mazur suite data54.9%
19GPT 5low reasoning· Lech Mazur suite data53.2%
20GPT 5.1medium reasoning· Lech Mazur suite data51.2%
21o3 Minihigh reasoning· Lech Mazur suite data49%
22GLM 4.7· Lech Mazur suite data47.2%
23Claude Opus 4.1thinking, 16k· Lech Mazur suite data45.3%
24Gemini 2.5 Pro· Lech Mazur suite data44.9%
25Kimi K2 Thinking64k· Lech Mazur suite data44.6%
26DeepSeek V4 Flashthinking· Lech Mazur suite data43.9%
27Gemini 2.5 Pro· Lech Mazur suite data40.1%
28o3 Minimedium reasoning· Lech Mazur suite data40.1%
29Qwen 3 235B A22B· Lech Mazur suite data39.8%
30Claude Opus 4thinking, 16k· Lech Mazur suite data35.2%
31Gemini 2.5 Pro Preview 05-06· Lech Mazur suite data28.1%
32Claude Sonnet 4thinking, 16k· Lech Mazur suite data26.4%
33Claude Sonnet 4thinking, 64k· Lech Mazur suite data26%
34DeepSeek R1· Lech Mazur suite data24.4%
35GPT OSS 120B· Lech Mazur suite data24.1%
36Claude Opus 4.1no reasoning· Lech Mazur suite data21.7%
37Qwen3 30B A3B· Lech Mazur suite data21.2%
38Qwen3 32B· Lech Mazur suite data21%
39Claude Opus 4no reasoning· Lech Mazur suite data19.7%
40Claude Sonnet 3.7thinking, 16k· Lech Mazur suite data19.7%
41GPT-4.5 Preview· Lech Mazur suite data19.2%
42Qwen3 Next 80B A3B· Lech Mazur suite data18.5%
43QwQ 32B16k· Lech Mazur suite data17.8%
44Grok 3 Minihigh, beta· Lech Mazur suite data17.2%
45GLM 4.5· Lech Mazur suite data17%
46O1 Mini· Lech Mazur suite data14.1%
47Grok 3 Minilow, beta· Lech Mazur suite data14%
48GPT 5minimal reasoning· Lech Mazur suite data13.9%
49Claude Sonnet 4no reasoning· Lech Mazur suite data13.7%
50Claude Opus 4.632k· Lech Mazur suite data13.6%

Showing 1–50 of 93 results.