--- title: Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost? description: A benchmark study testing frontier LLMs (Claude Fable 5, GPT-5.6 Sol, Gemini 3.5 Flash, GLM-5.2, and others) on the puzzle game Baba Is You, ported to the... canonical: https://daily.dev/posts/baba-is-solved-by-fable-5-and-gpt-5-6-sol-but-at-what-cost--nvg87ohj0 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost? | daily.dev og:description: A benchmark study testing frontier LLMs (Claude Fable 5, GPT-5.6 Sol, Gemini 3.5 Flash, GLM-5.2, and others) on the puzzle game Baba Is You, ported to the... og:url: https://daily.dev/posts/baba-is-solved-by-fable-5-and-gpt-5-6-sol-but-at-what-cost--nvg87ohj0 og:image: https://api.daily.dev/og/posts/nVG87OhJ0.png og:image:alt: Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost? 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. # Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost? **[Quesma](https://daily.dev/sources/quesma)** · 13 min read · 4 upvotes · 0 comments ## Summary A benchmark study testing frontier LLMs (Claude Fable 5, GPT-5.6 Sol, Gemini 3.5 Flash, GLM-5.2, and others) on the puzzle game Baba Is You, ported to the Harbor agent evaluation framework. Claude Fable 5 and GPT-5.6 Sol solved 13 of 14 levels in Stage 1, but both were 4x and 13x slower respectively than a human Twitch streamer playing for the first time. Key findings: token price is misleading — total cost per task matters more; Gemini 3.5 Flash was 2.4x more expensive than Fable 5 despite lower token prices due to excessive turns; GPT-5.6 Sol was the most cost-efficient frontier model; open-weight GLM-5.2 outperformed both Gemini models. The team spent over $2,000 running experiments. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How much faster is a human at solving Baba Is You puzzle levels compared to Claude Fable 5 and GPT-5.6 Sol? A Twitch streamer playing live for the first time solved the levels 4x faster than Claude Fable 5 and 13x faster than GPT-5.6 Sol, often finishing in under 3 minutes while the models spent tens of minutes per level. The streamer also solved a bonus level, Sunken Temple, that neither frontier model could beat. _Anyone weighing AI agents against human effort on reasoning tasks can track benchmarks like this on daily.dev._ ### Which LLM was most cost-efficient at solving Baba Is You puzzle levels in the Baba Is Harbor benchmark? GPT-5.6 Sol was the most cost-efficient, costing 47% less than Claude Fable 5 and even less than the open-weight GLM-5.2 model, despite GLM-5.2 having a 4.1x lower per-token price. GLM-5.2 ended up 1.7x more expensive overall because it generated far more tokens to reach a solution. _Developers comparing model pricing versus real task cost can follow this kind of analysis on daily.dev._ ### Why was Gemini 3.5 Flash so expensive to run on the Baba Is You benchmark despite being marketed as a budget model? Gemini 3.5 Flash cost 2.4x more than Claude Fable 5 to solve the Intro stage because it took an average of 111 turns per level, a trigger-happy approach compared to frontier models. Total spend on Gemini 3.5 Flash runs reached about $220, with the model building up 500k-800k token context windows without reaching correct solutions on harder levels. _Teams budgeting for agentic AI workloads can watch for cost surprises like this via daily.dev._ ## Similar posts on daily.dev - [Fable 5 Is Back — And the Numbers Explain Why Everyone’s Talking About It Again](https://daily.dev/posts/fable-5-is-back-and-the-numbers-explain-why-everyone-s-talking-about-it-again-rkodhshha) · Medium · 0 upvotes · 0 comments - [Fable 5 vs GPT-5.6 Sol: I spent $2,000 and 2 billion tokens to find out who wins — Expo blog](https://daily.dev/posts/fable-5-vs-gpt-5-6-sol-i-spent-2-000-and-2-billion-tokens-to-find-out-who-wins-expo-blog-hfefgd2ti) · Expo Blog · 0 upvotes · 0 comments - [GPT-5.6 Sol vs Claude Fable 5: Which Model Writes Better Plans?](https://daily.dev/posts/gpt-5-6-sol-vs-claude-fable-5-which-model-writes-better-plans--m7odzpdq7) · Kilo Blog · 1 upvotes · 0 comments --- Tags: [#ai](https://daily.dev/tags/ai), [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents) [View this post on daily.dev](https://daily.dev/posts/baba-is-solved-by-fable-5-and-gpt-5-6-sol-but-at-what-cost--nvg87ohj0) ```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":"Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost?","url":"https://daily.dev/posts/baba-is-solved-by-fable-5-and-gpt-5-6-sol-but-at-what-cost--nvg87ohj0","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/baba-is-solved-by-fable-5-and-gpt-5-6-sol-but-at-what-cost--nvg87ohj0"},"datePublished":"2026-07-16T14:44:15.271Z","dateModified":"2026-09-14T06:49:19.732Z","description":"A benchmark study testing frontier LLMs (Claude Fable 5, GPT-5.6 Sol, Gemini 3.5 Flash, GLM-5.2, and others) on the puzzle game Baba Is You, ported to the...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/b156d3e892324efa49d5a12d99a98a6c?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/b156d3e892324efa49d5a12d99a98a6c?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Quesma","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":"Quesma","logo":"https://media.daily.dev/image/upload/s--I-Be0YJY--/f_auto,q_auto/v1774964372/logos/quesma?_a=BAMAMiWQ0","url":"https://daily.dev/sources/quesma"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/baba-is-solved-by-fable-5-and-gpt-5-6-sol-but-at-what-cost--nvg87ohj0","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":4},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai,llm,ai-agents","timeRequired":"PT13M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Quesma","item":"https://daily.dev/sources/quesma"},{"@type":"ListItem","position":3,"name":"Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost?"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/baba-is-solved-by-fable-5-and-gpt-5-6-sol-but-at-what-cost--nvg87ohj0#faq","mainEntity":[{"@type":"Question","name":"How much faster is a human at solving Baba Is You puzzle levels compared to Claude Fable 5 and GPT-5.6 Sol?","acceptedAnswer":{"@type":"Answer","text":"A Twitch streamer playing live for the first time solved the levels 4x faster than Claude Fable 5 and 13x faster than GPT-5.6 Sol, often finishing in under 3 minutes while the models spent tens of minutes per level. The streamer also solved a bonus level, Sunken Temple, that neither frontier model could beat. Anyone weighing AI agents against human effort on reasoning tasks can track benchmarks like this on daily.dev."}},{"@type":"Question","name":"Which LLM was most cost-efficient at solving Baba Is You puzzle levels in the Baba Is Harbor benchmark?","acceptedAnswer":{"@type":"Answer","text":"GPT-5.6 Sol was the most cost-efficient, costing 47% less than Claude Fable 5 and even less than the open-weight GLM-5.2 model, despite GLM-5.2 having a 4.1x lower per-token price. GLM-5.2 ended up 1.7x more expensive overall because it generated far more tokens to reach a solution. Developers comparing model pricing versus real task cost can follow this kind of analysis on daily.dev."}},{"@type":"Question","name":"Why was Gemini 3.5 Flash so expensive to run on the Baba Is You benchmark despite being marketed as a budget model?","acceptedAnswer":{"@type":"Answer","text":"Gemini 3.5 Flash cost 2.4x more than Claude Fable 5 to solve the Intro stage because it took an average of 111 turns per level, a trigger-happy approach compared to frontier models. Total spend on Gemini 3.5 Flash runs reached about $220, with the model building up 500k-800k token context windows without reaching correct solutions on harder levels. Teams budgeting for agentic AI workloads can watch for cost surprises like this via daily.dev."}}]} ```