Ducky Bench (Saxo Frog): leaderboard
ELO-ranked SVG generation benchmark where models recreate a reference frog image as vector graphics, judged by human voters on visual accuracy.
Metric: ELO. Source: ducky-bench.joinity.site. Saturation forecast: Around January 2027. 25 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O3 (High) | 1449 | #124 (O3) |
| 2 | Claude Opus 4 (Thinking) | 1280 | #163 (Claude Opus 4) |
| 3 | GPT-5 (High) | 1261 | #98 (GPT-5) |
| 4 | GPT-4.1 | 1241 | #248 |
| 5 | Grok 4 | 1229 | #181 |
| 6 | Claude Sonnet 4 (Thinking) | 1223 | #199 (Claude Sonnet 4) |
| 7 | Gemini 2.5 Pro | 1170 | #152 |
| 8 | O3 Pro | 1158 | #116 |
| 9 | Claude 3.7 Sonnet (Thinking) | 1142 | #249 (Claude 3.7 Sonnet) |
| 10 | Claude Sonnet 4.5 | 1098 | #142 |
| 11 | GPT-4.5 | 1095 | #274 |
| 12 | Gemini 3 Pro | 1081 | #80 |
| 13 | Gemini 2.5 Flash | 1063 | #245 |
| 14 | GPT-5 Mini (High) | 995 | #186 (GPT-5 Mini) |
| 15 | Grok 3 | 981 | #304 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=ducky-bench-saxo-frog · How It Works · Data refreshed daily, snapshot 2026-10-20.