--- title: Learning to play Minecraft with Video PreTraining description: OpenAI introduces Video PreTraining (VPT), a semi-supervised imitation learning method that trains agents to act by watching unlabeled internet videos. A small... canonical: https://daily.dev/posts/learning-to-play-minecraft-with-video-pretraining-bzzj6iovr twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Learning to play Minecraft with Video PreTraining | daily.dev og:description: OpenAI introduces Video PreTraining (VPT), a semi-supervised imitation learning method that trains agents to act by watching unlabeled internet videos. A small... og:url: https://daily.dev/posts/learning-to-play-minecraft-with-video-pretraining-bzzj6iovr og:image: https://api.daily.dev/og/posts/BzzJ6IOVR.png og:image:alt: Learning to play Minecraft with Video PreTraining 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. # Learning to play Minecraft with Video PreTraining **[OpenAI](https://daily.dev/sources/openai)** · 8 min read · 0 upvotes · 0 comments ## Summary OpenAI introduces Video PreTraining (VPT), a semi-supervised imitation learning method that trains agents to act by watching unlabeled internet videos. A small labeled dataset from contractors is used to train an inverse dynamics model (IDM), which then labels 70,000 hours of Minecraft gameplay. The resulting foundation model learns complex behaviors like chopping trees, crafting tools, swimming, and hunting. Fine-tuned with reinforcement learning, the model becomes the first AI to craft diamond tools in Minecraft — a task requiring over 20 minutes and 24,000 actions for skilled humans. The approach uses the native mouse-and-keyboard interface, suggesting broad applicability to general computer-use agents. Code, model weights, and contractor data are open-sourced, and the work is tied to a NeurIPS 2022 competition with a $100,000 prize. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Similar posts on daily.dev - [A General Goal-Conditioned Minecraft Model](https://daily.dev/posts/a-general-goal-conditioned-minecraft-model-ydhsvavpp) · Hacker News · 0 upvotes · 0 comments - [Towards Data Science](https://daily.dev/posts/towards-data-science-n9ra6ei3l) · Towards Data Science · 0 upvotes · 0 comments --- Tags: [#ai](https://daily.dev/tags/ai), [#openai](https://daily.dev/tags/openai), [#computer-vision](https://daily.dev/tags/computer-vision), [#reinforcement-learning](https://daily.dev/tags/reinforcement-learning) [View this post on daily.dev](https://daily.dev/posts/learning-to-play-minecraft-with-video-pretraining-bzzj6iovr) ```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":"Learning to play Minecraft with Video PreTraining","url":"https://daily.dev/posts/learning-to-play-minecraft-with-video-pretraining-bzzj6iovr","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/learning-to-play-minecraft-with-video-pretraining-bzzj6iovr"},"datePublished":"2026-07-09T18:39:05.765Z","dateModified":"2026-07-09T19:27:39.119Z","description":"OpenAI introduces Video PreTraining (VPT), a semi-supervised imitation learning method that trains agents to act by watching unlabeled internet videos. A small...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/02f5c17e2f6383e495e65680431dfe74?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/02f5c17e2f6383e495e65680431dfe74?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"OpenAI","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":"OpenAI","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/984b3632316f4afaa3d5213d5c64fe65","url":"https://daily.dev/sources/openai"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/learning-to-play-minecraft-with-video-pretraining-bzzj6iovr","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai,openai,computer-vision,reinforcement-learning","timeRequired":"PT8M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"OpenAI","item":"https://daily.dev/sources/openai"},{"@type":"ListItem","position":3,"name":"Learning to play Minecraft with Video PreTraining"}]} ```