--- title: Claude Fable 5.1 made me a really nice animated pelican description: Simon Willison tests Anthropic's newly released Claude Fable 5.1 using his informal 'pelican riding a bicycle' SVG benchmark across all five reasoning effort... canonical: https://daily.dev/posts/claude-fable-5-1-made-me-a-really-nice-animated-pelican-dymfmpnkq twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Claude Fable 5.1 made me a really nice animated pelican | daily.dev og:description: Simon Willison tests Anthropic's newly released Claude Fable 5.1 using his informal 'pelican riding a bicycle' SVG benchmark across all five reasoning effort... og:url: https://daily.dev/posts/claude-fable-5-1-made-me-a-really-nice-animated-pelican-dymfmpnkq og:image: https://api.daily.dev/og/posts/dymfmpnkq.png og:image:alt: Claude Fable 5.1 made me a really nice animated pelican og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Claude Fable 5.1 made me a really nice animated pelican **[Simon Willison](https://daily.dev/sources/simonwillison)** · 6 min read · 0 upvotes · 0 comments ## Summary Simon Willison tests Anthropic's newly released Claude Fable 5.1 using his informal 'pelican riding a bicycle' SVG benchmark across all five reasoning effort levels (low, medium, high, xhigh, max). Low and medium showed no visible reasoning tokens and nearly identical output, high added minor reasoning, while xhigh and max produced dramatically more detailed pelicans at much higher cost ($1.83 and $3.30 respectively) and longer runtimes (up to nearly 14 minutes). He also generated an animated version of the max-effort pelican for $1.37 using the high reasoning tier. Willison notes he's losing confidence in the pelican benchmark's correlation with broader model capability, though it still reveals interesting differences across reasoning levels within a model family. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What are the reasoning effort levels available in Claude Fable 5.1? Claude Fable 5.1 offers five reasoning effort levels: low, medium, high, xhigh, and max, with no option to disable reasoning entirely. Low and medium produced nearly identical output (around 1,977-1,998 output tokens) with no visible reasoning trace, while xhigh jumped to 36,767 output tokens and max reached 65,927 output tokens, taking nearly 14 minutes and costing $3.30 for a single SVG generation prompt. _Developers weighing cost versus output quality across reasoning tiers can track model changes like this on daily.dev._ ### How did Claude Fable 5.1 score on the Terminal-Bench-Science 0.1 benchmark compared to previous Anthropic models? Claude Fable 5.1 scored 52.6% on the Terminal-Bench-Science 0.1 benchmark, a large jump from 24.7% for Fable 5 and 29.0% for Opus 5. For comparison, GPT-5.6 Sol scored 22.4% on the same benchmark, which was first announced on August 27th, 2026. _Anyone comparing new model releases against prior benchmarks can follow coding-model updates on daily.dev._ ### Why does increasing reasoning effort in Claude Fable 5.1 sometimes produce no visible reasoning trace? At the low and medium effort settings, Claude Fable 5.1 appeared to skip reasoning entirely for a simple SVG generation prompt, showing no summarized reasoning tokens and using roughly the same output token count (1,998 and 1,977 respectively). Meaningful reasoning traces only appeared starting at the high effort level, with a small reasoning summary, and became extensive at xhigh and max. _Developers debugging reasoning-token behavior across effort settings can dig into model release notes on daily.dev._ ## Similar posts on daily.dev - [Are AI labs pelicanmaxxing? – Dylan Castillo](https://daily.dev/posts/are-ai-labs-pelicanmaxxing-dylan-castillo-uc46wcun6) · Hacker News · 0 upvotes · 0 comments - [What happens if AI labs train for pelicans riding bicycles?](https://daily.dev/posts/what-happens-if-ai-labs-train-for-pelicans-riding-bicycles--lb0v74rks) · Simon Willison · 2 upvotes · 0 comments - [Agentic Pelican on a Bicycle](https://daily.dev/posts/agentic-pelican-on-a-bicycle-6v5vgmwem) · Hacker News · 1 upvotes · 0 comments --- Tags: [#ai](https://daily.dev/tags/ai), [#genai](https://daily.dev/tags/genai), [#claude](https://daily.dev/tags/claude), [#anthropic](https://daily.dev/tags/anthropic) [View this post on daily.dev](https://daily.dev/posts/claude-fable-5-1-made-me-a-really-nice-animated-pelican-dymfmpnkq) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Claude Fable 5.1 made me a really nice animated pelican","url":"https://daily.dev/posts/claude-fable-5-1-made-me-a-really-nice-animated-pelican-dymfmpnkq","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/claude-fable-5-1-made-me-a-really-nice-animated-pelican-dymfmpnkq"},"datePublished":"2026-09-02T00:01:39.191Z","dateModified":"2026-09-02T00:17:35.044Z","description":"Simon Willison tests Anthropic's newly released Claude Fable 5.1 using his informal 'pelican riding a bicycle' SVG benchmark across all five reasoning effort...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/df0cc6242fcaafd9c6da0d19158b1c83?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/df0cc6242fcaafd9c6da0d19158b1c83?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Simon Willison","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Simon Willison","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/cf12b897300740218b35f49a6309ec5b","url":"https://daily.dev/sources/simonwillison"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/claude-fable-5-1-made-me-a-really-nice-animated-pelican-dymfmpnkq","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai,genai,claude,anthropic","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Simon Willison","item":"https://daily.dev/sources/simonwillison"},{"@type":"ListItem","position":3,"name":"Claude Fable 5.1 made me a really nice animated pelican"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/claude-fable-5-1-made-me-a-really-nice-animated-pelican-dymfmpnkq#faq","mainEntity":[{"@type":"Question","name":"What are the reasoning effort levels available in Claude Fable 5.1?","acceptedAnswer":{"@type":"Answer","text":"Claude Fable 5.1 offers five reasoning effort levels: low, medium, high, xhigh, and max, with no option to disable reasoning entirely. Low and medium produced nearly identical output (around 1,977-1,998 output tokens) with no visible reasoning trace, while xhigh jumped to 36,767 output tokens and max reached 65,927 output tokens, taking nearly 14 minutes and costing $3.30 for a single SVG generation prompt. Developers weighing cost versus output quality across reasoning tiers can track model changes like this on daily.dev."}},{"@type":"Question","name":"How did Claude Fable 5.1 score on the Terminal-Bench-Science 0.1 benchmark compared to previous Anthropic models?","acceptedAnswer":{"@type":"Answer","text":"Claude Fable 5.1 scored 52.6% on the Terminal-Bench-Science 0.1 benchmark, a large jump from 24.7% for Fable 5 and 29.0% for Opus 5. For comparison, GPT-5.6 Sol scored 22.4% on the same benchmark, which was first announced on August 27th, 2026. Anyone comparing new model releases against prior benchmarks can follow coding-model updates on daily.dev."}},{"@type":"Question","name":"Why does increasing reasoning effort in Claude Fable 5.1 sometimes produce no visible reasoning trace?","acceptedAnswer":{"@type":"Answer","text":"At the low and medium effort settings, Claude Fable 5.1 appeared to skip reasoning entirely for a simple SVG generation prompt, showing no summarized reasoning tokens and using roughly the same output token count (1,998 and 1,977 respectively). Meaningful reasoning traces only appeared starting at the high effort level, with a small reasoning summary, and became extensive at xhigh and max. Developers debugging reasoning-token behavior across effort settings can dig into model release notes on daily.dev."}}]} ```