Skip to content

Benchmark

SVGBench

Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.

Models measured
82
Scores recorded
88
Sources
3
IQ weight
2.21

Score distribution

47.1%
Mean
48.2%
Median
14.6%
Std dev
28.4%
10th percentile
64.1%
90th percentile
6.8%
Min
6.8%74.4%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.21 for chat models, measured over 82 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1GPT 5.2svg-generation, xhigh· SVGBench data74.4%
2Claude Opus 4.5svg-generation, non-thinking· SVGBench data72%
3Claude Opus 4.5svg-generation, thinking· SVGBench data71.5%
4GLM 5svg-generation· SVGBench data70.3%
5Gemini 3 Pro Previewsvg-generation· SVGBench data68.7%
6GPT 5.1svg-generation, high· SVGBench data67.5%
7Kimi K2.5svg-generation· SVGBench data67.5%
8GPT 5svg-generation, high· SVGBench data67.4%
9MiniMax M2.5svg-generation· SVGBench data65%
10Gemini 3 Flashsvg-generation· SVGBench data64.1%
11MiniMax M3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim63.7%
12MiniMax M3official model-page overview figure· Reported by MiniMax· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.63.7%
13MiniMax M3text/image inputs · VLM rendering verifier · three-run mean· Reported by MiniMax Research· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.63.7%
14MiniMax M2.1svg-generation· SVGBench data63.1%
15GLM 4.7svg-generation· SVGBench data62.4%
16Claude Sonnet 4.5svg-generation· SVGBench data62.2%
17Gemini 2.5 Prosvg-generation· SVGBench data61.4%
18GPT-5 Codexsvg-generation, medium· SVGBench data61%
19GLM 4.6svg-generation· SVGBench data60.5%
20Claude Sonnet 3.7svg-generation· SVGBench data60.4%
21GPT-5 minisvg-generation· SVGBench data59.7%
22Claude Sonnet 4svg-generation· SVGBench data59.4%
23GPT 4.1svg-generation· SVGBench data58.4%
24GLM 4.5svg-generation· SVGBench data58.3%
25MiniMax M2svg-generation· SVGBench data58.3%
26Claude Haiku 4.5svg-generation· SVGBench data57.2%
27Kimi K2 Thinkingsvg-generation· SVGBench data57.1%
28OpenAI o3svg-generation· SVGBench data56.7%
29Grok 4svg-generation· SVGBench data54.4%
30GPT-4.1 minisvg-generation· SVGBench data53.4%
31DeepSeek V3.1svg-generation· SVGBench data53.1%
32gemini-2.5-flash-preview-09-2025svg-generation· SVGBench data52.1%
33Gemini 2.5 Flashsvg-generation· SVGBench data51.4%
34DeepSeek V3.2 Expsvg-generation· SVGBench data51.3%
35Grok 4 Fastsvg-generation· SVGBench data50.5%
36GPT-5 Chatsvg-generation· SVGBench data50.4%
37DeepSeek V3.2svg-generation· SVGBench data49.8%
38Qwen3 Coder Plussvg-generation· SVGBench data49.8%
39GPT-5 nanosvg-generation· SVGBench data49.7%
40GLM 4.5 Airsvg-generation· SVGBench data48.9%
41Qwen3 Codersvg-generation· SVGBench data48.7%
42Qwen3 235B A22B Instruct 2507svg-generation· SVGBench data48.6%
43Kimi K2 0905svg-generation· SVGBench data48.5%
44DeepSeek V3.1 Terminussvg-generation, reasoning· SVGBench data48.2%
45o4 Minisvg-generation· SVGBench data48%
46Claude 3.5 Sonnetsvg-generation· SVGBench data47.9%
47DeepSeek V3.1svg-generation, reasoning· SVGBench data47.8%
48Qwen3 Next 80B A3B Instructsvg-generation· SVGBench data47.4%
49GLM 4.5Vsvg-generation· SVGBench data46.5%
50GPT-5.1 Chatsvg-generation· SVGBench data46.3%

Showing 1–50 of 88 results.

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Data by LLM Stats