VectorGym - Text2SVG: leaderboard
Metric: Text2SVG task score (0-100): SVG code from a human-written description, mean of the Gemini 3 Pro judge score, CLIP and DINO similarity (all 0-100), on VectorGym's human-annotated test split (real-world SVGs from SVG-Stack); higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.1 | 70.68 | #131 |
| 2 | Gemini 3 Pro | 69.92 | #77 |
| 3 | Claude Sonnet 4.5 | 69.19 | #138 |
| 4 | GPT-4o | 63.09 | #333 |
| 5 | Qwen 3 VL 235B A22B Instruct | 59.68 | #264 |
| 6 | GLM-4.5V | 56.25 | #339 |
| 7 | Qwen 3 VL 8B Instruct | 55.46 | #401 |
| 8 | Gemini 2.5 Flash | 53.11 | #237 |
| 9 | Qwen 2.5 VL 72B Instruct | 40.86 | #364 |
| 10 | Qwen 2.5 VL 32B Instruct | 38.84 | #443 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=vectorgym-text2svg · How It Works · Data refreshed daily, snapshot 2026-10-11.