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
101Doubao-Seed 1.6· SnakeBench data24.9%
102MiniMax M3· SnakeBench data24.8%
103Devstral 2 123b Instruct 2512· SnakeBench data24.7%
104Kimi K2.5· SnakeBench data24.7%
105Gemma 3 27B IT· SnakeBench data24.6%
106GPT-5.6 Sol· SnakeBench data24.6%
107Nemotron 3 Ultra· SnakeBench data24.2%
108MiniMax M1· SnakeBench data24.1%
109Qwen3.6 35B A3B· SnakeBench data24%
110MiniMax M2· SnakeBench data24%
111Qwen3 32B· SnakeBench data23.9%
112Hy3 (Deep Infra)· SnakeBench data23.9%
113GLM 4 32B· SnakeBench data23.8%
114GLM 4.5 Air· SnakeBench data23.7%
115Gemma 4 26B A4B· SnakeBench data23.7%
116o4 Mini Deep Research· SnakeBench data23.6%
117Seed 2.0 Lite· SnakeBench data23.5%
118Qwen3.7 Plus· SnakeBench data23.4%
119Step 3.5 Flash· SnakeBench data23.4%
120DeepSeek V3.1· SnakeBench data23.4%
121Gemini 2.5 Flash Lite Preview 09-2025· SnakeBench data23.3%
122GPT-5.1 Chat· SnakeBench data23.3%
123Aion-1.0· SnakeBench data23.2%
124ERNIE 4.5 300B A47B· SnakeBench data23.1%
125GPT-4o Search Preview· SnakeBench data23.1%
126Cydonia 24B V4.1· SnakeBench data23%
127Llama 4 Maverick· SnakeBench data22.9%
128Aion-1.0-Mini· SnakeBench data22.8%
129Seed 2.0 Mini· SnakeBench data22.8%
130Ring 2.6 1T· SnakeBench data22.7%
131Qwen3 Coder 30B A3B Instruct· SnakeBench data22.6%
132Qwen Plusthinking· SnakeBench data22.6%
133Qwen3.5 397B A17B· SnakeBench data22.5%
134MiMo V2 Flash· SnakeBench data22.5%
135Aion 2.0· SnakeBench data22%
136MiniMax M2.7· SnakeBench data21.7%
137Olmo 3.1 32B Think· SnakeBench data21.6%
138Mistral 7B Instruct· SnakeBench data21.4%
139Gemini 2.0 Flash Lite· SnakeBench data21.3%
140Qwen3 30B A3B Instruct· SnakeBench data21.3%
141Llama 3.3 70B Instruct· SnakeBench data21.3%
142MiMo V2.5· SnakeBench data21.1%
143GPT-5.3 Chat· SnakeBench data21%
144Step 3.7 Flash· SnakeBench data21%
145DeepSeek V3.1 Terminus· SnakeBench data20.8%
146GLM 4.5V· SnakeBench data20.8%
147Palmyra X5· SnakeBench data20.8%
148Pixtral Large2411· SnakeBench data20.8%
149Qwen3 Coder Next· SnakeBench data20.7%
150MiMo V2 Pro· SnakeBench data20.7%

Showing 101–150 of 267 results.

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