Skip to content

Benchmark

Debate (Lech Mazur)

Models measured
64
Scores recorded
67
Sources
1
IQ weight
2.13

Score distribution

1,501.3
Mean
1,520.3
Median
139.9
Std dev
1,349.7
10th percentile
1,676.2
90th percentile
1,006.3
Min
1,006.31,705.5
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.13 for chat models, measured over 64 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 Fable 5.1high· Lech Mazur suite data1,705.5
2Claude Fable 5high· Lech Mazur suite data1,705.2
3Claude Opus 5high· Lech Mazur suite data1,703.7
4Claude Fable 5.2high· Lech Mazur suite data1,687.9
5Claude Opus 5.5high· Lech Mazur suite data1,680.3
6Qwen4 Maxhigh· Lech Mazur suite data1,678.8
7Kimi K3· Lech Mazur suite data1,678.6
8Gemini 4 Carbonhigh· Lech Mazur suite data1,676.2
9GPT-6.2 Astrahigh· Lech Mazur suite data1,673.2
10Gemini 4 Argonhigh· Lech Mazur suite data1,649
11Hy 4high· Lech Mazur suite data1,646.5
12Muse Wonder 1.0high· Lech Mazur suite data1,637.2
13Claude Opus 4.7high· Lech Mazur suite data1,627.4
14GPT-5.6 Solhigh· Lech Mazur suite data1,624.7
15Muse Spark 1.1high· Lech Mazur suite data1,617
16GLM 5.3high· Lech Mazur suite data1,613.1
17Claude Opus 4.8high· Lech Mazur suite data1,610.3
18GPT-6 Astrahigh· Lech Mazur suite data1,606.6
19Grok 4.7high· Lech Mazur suite data1,601.1
20Kimi K3.1 Prohigh· Lech Mazur suite data1,601
21Grok 4.6high· Lech Mazur suite data1,585
22MiMo V2.6 Prothinking· Lech Mazur suite data1,584
23DeepSeek V4.1 Flash 1017high· Lech Mazur suite data1,578.7
24Muse Spark 1.3high· Lech Mazur suite data1,577.6
25Claude Sonnet 5high· Lech Mazur suite data1,567.7
26DeepSeek V4.1 Flashhigh· Lech Mazur suite data1,564.2
27Hy4 previewhigh· Lech Mazur suite data1,561.3
28DeepSeek V4 Pro 0813high· Lech Mazur suite data1,552.3
29GLM 5.2max· Lech Mazur suite data1,542.3
30Claude Sonnet 4.6high· Lech Mazur suite data1,539.6
31Qwen3.8 Max· Lech Mazur suite data1,534.8
32GPT 5.4high· Lech Mazur suite data1,525.4
33Step 5 Previewhigh· Lech Mazur suite data1,521.1
34Claude Sonnet 4.6no reasoning· Lech Mazur suite data1,520.3
35MiniMax M3.1 Flashhigh· Lech Mazur suite data1,509.4
36GPT 5.5high· Lech Mazur suite data1,506.8
37GLM 5.1· Lech Mazur suite data1,497.7
38Kimi K2.6· Lech Mazur suite data1,490.4
39Gemini 3.8 Flashhigh· Lech Mazur suite data1,487.9
40MiniMax M3· Lech Mazur suite data1,482.9
41GPT 5.4no reasoning· Lech Mazur suite data1,472.8
42Grok 4.5high· Lech Mazur suite data1,469.9
43Gemini 3.1 Pro Preview· Lech Mazur suite data1,468.2
44MiMo V2.5 Pro· Lech Mazur suite data1,463
45Qwen3.6 Max Preview· Lech Mazur suite data1,453.5
46Mistral Large 4high· Lech Mazur suite data1,442.3
47Gemini 3.7 Flashhigh· Lech Mazur suite data1,434.2
48Kimi K2.5thinking· Lech Mazur suite data1,431.2
49DeepSeek-V4-Pro Preview· Lech Mazur suite data1,427.1
50Qwen3.7 Max· Lech Mazur suite data1,413.6

Showing 1–50 of 67 results.