Testing Claude Fable 5.1: Max mode produces the best pelican SVG for $3.30
Simon Willison · rss · 2026-09-02
Simon Willison tested the newly released Claude Fable 5.1, focusing on its ability and cost to generate an SVG of a "pelican riding a bicycle" across different reasoning effort levels.
- Low/Medium: Skipped reasoning almost entirely, costing $0.10 in 23s.
- High: Showed minimal reasoning summary, costing $0.13 in 29.6s.
- XHigh: Extensive reasoning trace with detailed adjustments (e.g., resizing hats, adding baskets), costing $1.83 in 7m 51s.
- Max: Produced the best result with background, basket, and blue hat, costing $3.30 in 13m 54s.
The author then piped the static Max SVG back into the model with the prompt "animate this" to successfully convert it into an animated SVG.
More from Models
- OpenAI's 'recurrent depth' reasoning approach raises monitoring concerns — steph_palazzolo · 2026-09-02
- Fable 5.1 achieves better results at lower cost on low-effort settings — sven_ai · 2026-09-02
- OpenAI explores looped transformers to improve answers by reprocessing text — pstAsiatech · 2026-09-02
- Claude Fable 5.1 preserves prompt cache when adjusting thinking effort — brandon_galang · 2026-09-02
- Anthropic launches Mythos 5.1 with Life Sciences Verification Program — arjunrajlab · 2026-09-02
- Dev: Anthropic 5.1 is a massive leap over previous model — doodlestein · 2026-09-02