Benchmark
SnakeBench
Scores are percentages.
All reasoning benchmarksScores from SnakeBench data.
- 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
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
| # | Model | Score |
|---|---|---|
| 251 | Phi-3.5 Mini 128K Instruct· SnakeBench data | 10% |
| 252 | Phi 4 Mini· SnakeBench data | 9.6% |
| 253 | DeepSeek V3.2 Speciale· SnakeBench data | 9.1% |
| 254 | Nova Micro 1.0· SnakeBench data | 9.1% |
| 255 | Ministral 3B· SnakeBench data | 9% |
| 256 | Llama 3.2 1B Instruct· SnakeBench data | 9% |
| 257 | DeepSeek V4 Pro· SnakeBench data | 8.6% |
| 258 | Relace Search· SnakeBench data | 8.5% |
| 259 | Morph V3 Fast· SnakeBench data | 8.5% |
| 260 | GPT 4o· SnakeBench data | 8.5% |
| 261 | Qwen2.5-VL 7B Instruct· SnakeBench data | 8.3% |
| 262 | GPT Audio Mini· SnakeBench data | 8.1% |
| 263 | Gemma 3 4B· SnakeBench data | 8.1% |
| 264 | Spotlight· SnakeBench data | 7.9% |
| 265 | Mistral (7B) Instruct v0.1· SnakeBench data | 7.6% |
| 266 | Reka Edge· SnakeBench data | 4.4% |
| 267 | Ling 2.6 Flash· SnakeBench data | 1.2% |
Showing 251–267 of 267 results.
Data credits: SnakeBench (Greg Kamradt / ARC Prize Foundation), MIT (licence)