Skip to content

Benchmark

Chess Puzzles

Scores are percentages.

Models measured
93
Scores recorded
163
Sources
1
IQ weight
1.82

Score distribution

29.5%
Mean
27%
Median
19.6%
Std dev
5%
10th percentile
59%
90th percentile
0%
Min
0%83%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.82 for chat models, measured over 93 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 Neonxhigh· Epoch AI Benchmarking Hub data83%
2Claude Fable 5.2max· Epoch AI Benchmarking Hub data79%
3Gemini 4 Carbonhigh· Epoch AI Benchmarking Hub data78%
4GPT-6.2 Astramax· Epoch AI Benchmarking Hub data78%
5GPT-6 Astrapro / max· Epoch AI Benchmarking Hub data76%
6GPT-6 Astramax· Epoch AI Benchmarking Hub data74%
7GPT-5.6 Solpro / max· Epoch AI Benchmarking Hub data68%
8GPT-6.1 Solmax· Epoch AI Benchmarking Hub data68%
9Claude Opus 5.5max· Epoch AI Benchmarking Hub data67%
10Gemini 3.7 Flashhigh· Epoch AI Benchmarking Hub data66%
11GPT 5.5xhigh· Epoch AI Benchmarking Hub data63%
12GPT-5.5 Proxhigh· Epoch AI Benchmarking Hub data62%
13Gemini 4 Hydrogenxhigh· Epoch AI Benchmarking Hub data60%
14GPT-5.2 Proxhigh· Epoch AI Benchmarking Hub data60%
15Claude Sonnet 5.5max· Epoch AI Benchmarking Hub data59%
16Gemini 4 Argonhigh· Epoch AI Benchmarking Hub data59%
17GPT-5.4 Proxhigh· Epoch AI Benchmarking Hub data59%
18GPT-5.6 Solmax· Epoch AI Benchmarking Hub data59%
19Gemini 3.8 Flashhigh· Epoch AI Benchmarking Hub data58%
20GPT-6 Solmax· Epoch AI Benchmarking Hub data56%
21Claude Haiku 5.5max· Epoch AI Benchmarking Hub data55%
22Claude Opus 4.8max· Epoch AI Benchmarking Hub data52%
23Gemini 3.1 Pro Previewhigh· Epoch AI Benchmarking Hub data51%
24Claude Fable 5.1max· Epoch AI Benchmarking Hub data48%
25DeepSeek V4.1 Flashmax· Epoch AI Benchmarking Hub data48%
26GPT 5.2xhigh· Epoch AI Benchmarking Hub data48%
27GPT 5.4xhigh· Epoch AI Benchmarking Hub data48%
28GPT-5 Prohigh· Epoch AI Benchmarking Hub data48%
29Claude Fable 5max· Epoch AI Benchmarking Hub data47%
30GPT-5.6 Terramax· Epoch AI Benchmarking Hub data47%
31DeepSeek V4 Pro 0813max· Epoch AI Benchmarking Hub data46%
32DeepSeek V4.1 Flash 1017max· Epoch AI Benchmarking Hub data46%
33Qwen3.8 Maxxhigh· Epoch AI Benchmarking Hub data46%
34Claude Fable 5high· Epoch AI Benchmarking Hub data45%
35Claude Sonnet 5max· Epoch AI Benchmarking Hub data45%
36Gemini 3.5 Flashhigh· Epoch AI Benchmarking Hub data45%
37GPT 5.2high· Epoch AI Benchmarking Hub data45%
38GPT 5.4 Nanoxhigh· Epoch AI Benchmarking Hub data45%
39Claude Opus 5max· Epoch AI Benchmarking Hub data43%
40GPT-5.6 Lunamax· Epoch AI Benchmarking Hub data43%
41Gemini 3.5 Flashlow· Epoch AI Benchmarking Hub data42%
42Gemini 3.6 Flashlow· Epoch AI Benchmarking Hub data42%
43GPT 5.1high· Epoch AI Benchmarking Hub data42%
44GPT 5.4high· Epoch AI Benchmarking Hub data42%
45Claude Opus 4.7max· Epoch AI Benchmarking Hub data41%
46Grok 4.7xhigh· Epoch AI Benchmarking Hub data41%
47Doubao Seed Evolvinghigh· Epoch AI Benchmarking Hub data40%
48Gemini 3 Flashhigh· Epoch AI Benchmarking Hub data40%
49Gemini 3.6 Flashhigh· Epoch AI Benchmarking Hub data40%
50GLM 5.3max· Epoch AI Benchmarking Hub data37%

Showing 1–50 of 163 results.

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