Simon Willison · 2026-09-04 · notable
Simon Willison — GPT-6 Astra draws far better pelicans than GPT-5.6
Simon Willison ran his pelican-on-a-bicycle SVG test on GPT-6 Astra at five reasoning levels and lined the results up against GPT-5.6 Sol, Terra and Luna. Astra's drawings are much better, and its cheapest run cost 9.55 cents.
A side-by-side grid puts GPT-6 Astra against three GPT-5.6 models at every reasoning level.
What is it?
Simon Willison published a comparison grid of SVG pelicans riding bicycles, drawn by GPT-6 Astra at low, medium, high, xhigh and max reasoning effort, next to GPT-5.6 Sol, Terra and Luna across their own effort levels. The pelican prompt is the informal test Willison runs on every new model.
How does it work?
Every cell of the grid records the input tokens, output tokens and cash cost of that one drawing, priced from the published rates in the page header: $10/$50 per million input/output tokens for Astra, $4/$20 for Sol, $2/$12 for Terra and $0.20/$1.20 for Luna. At max effort GPT-6 Astra spent 16 input and 12,638 output tokens, or 63.21 cents; Luna spent 13,040 output tokens for 1.57 cents.
Why does it matter?
Reasoning effort moves the bill as much as the model does — one pelican ranges from 1.57 to 63.21 cents across this grid. Willison writes that GPT-6 Astra at low effort already beats every GPT-5.6 Sol drawing for 9.55 cents. He also spotted that Astra and Luna both used 16 input tokens while Sol and Terra used 26, and speculates the two may be more closely related than OpenAI has said.
Who is it for?
developers choosing a model and a reasoning level
Try it
https://static.simonwillison.net/static/2026/gpt-6-and-5.6-pelicans.html