Skip to content

Benchmark

SnakeBench

Scores are percentages.

Models measured
260
Scores recorded
267
Sources
1
IQ weight
1.32

Score distribution

22.1%
Mean
22.5%
Median
7.3%
Std dev
13.2%
10th percentile
32.2%
90th percentile
1.2%
Min
1.2%38.5%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.32 for chat models, measured over 254 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 Image Mini· SnakeBench data28.7%
52mercury-coder· SnakeBench data28.6%
53Qwen3 VL 235B· SnakeBench data28.5%
54Gemini 2.5 Pro· SnakeBench data28.3%
55GLM 5V Turbo· SnakeBench data28.1%
56Mercury 2· SnakeBench data28.1%
57Codex Mini· SnakeBench data28.1%
58DeepSeek V4 Flash· SnakeBench data28%
59Qwen3.6 Plus· SnakeBench data27.9%
60Grok Code Fast 1· SnakeBench data27.9%
61GPT-5 nano· SnakeBench data27.8%
62Qwen-VL Max· SnakeBench data27.6%
63Mistral Small 3.2 24B Instruct· SnakeBench data27.4%
64Qwen3 Next 80B A3B Instruct· SnakeBench data27%
65KAT-Coder-Pro V2· SnakeBench data27%
66MiMo V2.5 Pro· SnakeBench data26.8%
67MiniMax M2.5· SnakeBench data26.6%
68Claude Sonnet 3.7thinking· SnakeBench data26.6%
69Qwen3.5 Plus· SnakeBench data26.5%
70Claude Sonnet 4.6· SnakeBench data26.4%
71GLM 5 Turbo· SnakeBench data26.3%
72Qwen3 Next 80B A3Bthinking· SnakeBench data26.3%
73Qwen Turbo· SnakeBench data26.3%
74GPT-5 Chat· SnakeBench data26.2%
75Qwen3 30B A3B2507, thinking· SnakeBench data26.2%
76GLM 4.6v· SnakeBench data26.1%
77GLM 5.1· SnakeBench data26%
78o3 Mini· SnakeBench data26%
79GLM 4.7· SnakeBench data26%
80GPT OSS 120B· SnakeBench data25.9%
81Magistral Medium2506· SnakeBench data25.8%
82Llama 3.1 Nemotron 70B Instruct· SnakeBench data25.8%
83Sonar Prosearch· SnakeBench data25.7%
84Tongyi DeepResearch 30B A3B· SnakeBench data25.7%
85GPT-5.2 Chat· SnakeBench data25.6%
86Claude Haiku 4.5· SnakeBench data25.6%
87GPT-5 mini· SnakeBench data25.6%
88ERNIE 4.5 VL 424B A47B· SnakeBench data25.3%
89Olmo 3 7B Think· SnakeBench data25.2%
90GLM 4.6· SnakeBench data25.2%
91Qwen3.5 35B A3B· SnakeBench data25.2%
92Gemini 2.5 Flash-Lite· SnakeBench data25.1%
93Claude Sonnet 3.7· SnakeBench data25%
94Qwen3 8B· SnakeBench data25%
95ling-2.6-1t· SnakeBench data25%
96Mistral-Small-3· SnakeBench data24.9%
97Llama 3.2 90B Vision Instruct· SnakeBench data24.9%
98Qwen3 Coder Plus· SnakeBench data24.9%
99Qwen3 VL 235Bthinking· SnakeBench data24.9%
100GPT-4.1 mini· SnakeBench data24.9%

Showing 51–100 of 267 results.

Data credits: SnakeBench (Greg Kamradt / ARC Prize Foundation), MIT (licence)