gemini-exp-1121
An early experimental model; it later described its own image as an insect.

Simon Willison tested Anthropic's newly released Claude Haiku 5.5 in his blog post 'Claude Haiku 5.5' using the llm CLI. The model defaults to medium reasoning effort and cannot disable thinking. The author tested effort levels from low to max and featured the max-effort output in the article, which took 5 minutes 9 seconds and cost ~3.38 cents. The output depicts a white pelican with a red cap, curved beak, long orange legs on pedals, dark grey bike frame, sun and speed streaks. Archives the article's original WebP preview from static.simonwillison.net; model label is source-reported and unauthenticated; date is the source publication on 2026-10-07.
Date precision / evidence: day / source-publication. Source publication: 2026-10-07. Generation date is not independently authenticated.
An early experimental model; it later described its own image as an insect.
An early locally generated pelican from an open reasoning model.
An early result the author first described as “spectacular.”