Skip to content

Benchmark

WeirdML

Scores are percentages.

Models measured
129
Scores recorded
164
Sources
1
IQ weight
2.56

Score distribution

49.8%
Mean
46.1%
Median
21.7%
Std dev
22.2%
10th percentile
82.9%
90th percentile
1.7%
Min
1.7%93.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.56 for chat models, measured over 129 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51GPT-5.4 minihigh· Epoch AI Benchmarking Hub data60.3%
52O3 Prohigh· Epoch AI Benchmarking Hub data58.2%
53GPT 5.4none· Epoch AI Benchmarking Hub data57.4%
54GPT-5.4 Pronone· Epoch AI Benchmarking Hub data57.4%
55GLM 5.1· Epoch AI Benchmarking Hub data57.1%
56DeepSeek V4 Flash 0731high· Epoch AI Benchmarking Hub data57.1%
57Gemini 3.6 Flashhigh· Epoch AI Benchmarking Hub data56.1%
58Kimi K2.6· Epoch AI Benchmarking Hub data55.9%
59GPT-5 Codex· Epoch AI Benchmarking Hub data54.5%
60Kimi K2.7 Code· Epoch AI Benchmarking Hub data54.1%
61Gemini 2.5 Pro16k· Epoch AI Benchmarking Hub data54%
62GPT-5 minihigh· Epoch AI Benchmarking Hub data52.7%
63o4 Minihigh· Epoch AI Benchmarking Hub data52.6%
64OpenAI o3high· Epoch AI Benchmarking Hub data52.4%
65Gemma 4 31B IT· Epoch AI Benchmarking Hub data52.3%
66Grok 4.20· Epoch AI Benchmarking Hub data52.3%
67Gemini 3.1 Flash Lite· Epoch AI Benchmarking Hub data52.2%
68Grok 4.3· Epoch AI Benchmarking Hub data49.9%
69GPT 5.2low· Epoch AI Benchmarking Hub data49.6%
70GPT 5.2none· Epoch AI Benchmarking Hub data49.6%
71GPT 5.4 Nanohigh· Epoch AI Benchmarking Hub data49.2%
72DeepSeek V4 Promax· Epoch AI Benchmarking Hub data48.6%
73GLM 5· Epoch AI Benchmarking Hub data48.2%
74GPT OSS 120Bhigh· Epoch AI Benchmarking Hub data48.2%
75Claude Sonnet 4.516k· Epoch AI Benchmarking Hub data47.7%
76O1 Preview· Epoch AI Benchmarking Hub data47.6%
77DeepSeek V3.2 Speciale· Epoch AI Benchmarking Hub data46.7%
78Claude Sonnet 4.5· Epoch AI Benchmarking Hub data46.7%
79DeepSeek V4 Prohigh· Epoch AI Benchmarking Hub data46.5%
80Grok 4.5· Epoch AI Benchmarking Hub data46.4%
81Claude Sonnet 416k· Epoch AI Benchmarking Hub data46.1%
82o1high· Epoch AI Benchmarking Hub data46.1%
83Claude Opus 4.116k· Epoch AI Benchmarking Hub data45.9%
84Grok 4· Epoch AI Benchmarking Hub data45.7%
85DeepSeek V4 Flashmax· Epoch AI Benchmarking Hub data45.6%
86Kimi K2.5· Epoch AI Benchmarking Hub data45.6%
87Claude Haiku 4.5· Epoch AI Benchmarking Hub data45.4%
88Claude Haiku 4.516k· Epoch AI Benchmarking Hub data44.2%
89Claude Sonnet 4· Epoch AI Benchmarking Hub data43.9%
90DeepSeek V4 Flashhigh· Epoch AI Benchmarking Hub data43.8%
91o3 Minihigh· Epoch AI Benchmarking Hub data43.8%
92Claude Opus 416k· Epoch AI Benchmarking Hub data43.7%
93Mistral Medium2604· Epoch AI Benchmarking Hub data43.7%
94Nemotron 3 Ultra· Epoch AI Benchmarking Hub data43.5%
95Mercury 2· Epoch AI Benchmarking Hub data43.2%
96Grok 4 Fast· Epoch AI Benchmarking Hub data42.9%
97Kimi K2 Thinking· Epoch AI Benchmarking Hub data42.8%
98Grok 3 Minihigh· Epoch AI Benchmarking Hub data42.6%
99Grok 3 Minihigh, beta· Epoch AI Benchmarking Hub data42.6%
100gemini-2.5-flash-preview-09-202516k· Epoch AI Benchmarking Hub data41.9%

Showing 51–100 of 164 results.

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