Simon Willison 11/25/2025

LLM SVG Generation Benchmark

Read Original

This article presents a benchmark project by Tom Gally, inspired by a previous SVG benchmark. It tests 9 frontier AI models on 30 creative text-to-SVG generation prompts, such as 'a starfish driving a bulldozer' and 'butterfly inspecting a steam engine'. The results provide a clear comparison of which large language models perform best at generating SVG artwork from imaginative descriptions.

LLM SVG Generation Benchmark

Comments

No comments yet

Be the first to share your thoughts!