Developer Retests LLMs on Creative SVG Prompts, Finds Stronger Results in 2026
A developer revisited a benchmark experiment originally inspired by Simon Willison's 'pelican riding a bicycle' test, in which large language models are prompted to generate creative SVGs. The original tests were conducted in November and December 2025 using around thirty imaginative prompts, such as an octopus playing a pipe organ. Nine months later, the developer reran ten of those prompts across six models via OpenRouter to compare how newer, more capable models perform. The updated results have been published on a dedicated website, with the experiment costing approximately twenty dollars before the developer decided to pause further testing.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)
Log in to join the discussion and vote.
Log in