{"id":6174,"date":"2026-10-19T16:49:47","date_gmt":"2026-10-19T16:49:47","guid":{"rendered":"https:\/\/i10xnews.kinsta.cloud\/?p=6174"},"modified":"2026-10-19T17:25:14","modified_gmt":"2026-10-19T17:25:14","slug":"spatialblock-synthetic-block-stacking-spatial-intelligence","status":"publish","type":"post","link":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence","title":{"rendered":"Teaching LVLMs Spatial Skills With 15K Toy Blocks Beats Heavier Real-Scene Pipelines"},"content":{"rendered":"\n<h2 class=\"wp-block-heading\">\u26a1 Quick Take<\/h2>\n\n<ul class=\"wp-block-list\"><li><strong>The finding:<\/strong> Spatial intelligence \u2014 reconstructing 3D structure from 2D images \u2014 remains weak in LVLMs. Humans in the paper&#8217;s teaser hit 95%, while GPT-5 and Gemini2.5-flash lag far behind on the illustrated item.<\/li><li><strong>Synthetic curriculum:<\/strong> SpatialBlock-15k offers 15,000 block-stacking problems covering projection, viewpoint transforms and structural combination, with controlled color cues.<\/li><li><strong>Transfer:<\/strong> A 4B model reaches 51.3% (+25.1 over backbone), best open-source in the reported comparison; 7B gains 17.6% over its backbone with only 15K samples.<\/li><\/ul>\n\n\n<p class=\"wp-block-paragraph\">Real-scene spatial QA needs expensive geometric labels. KAIST and AITRICS researchers take a developmental psychology cue: learn spatial foundations on structured block tasks first.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1344\" height=\"768\" src=\"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-gains.webp\" alt=\"Column chart: 2.7 point gain for 3B versus SpatialLadder, 17.6 for 7B over backbone, 25.1 for 4B over backbone.\" class=\"wp-image-6173\" srcset=\"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-gains.webp 1344w, https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-gains-300x171.webp 300w, https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-gains-1024x585.webp 1024w, https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-gains-768x439.webp 768w\" sizes=\"auto, (max-width: 1344px) 100vw, 1344px\" \/><figcaption class=\"wp-element-caption\">i10X chart based on data from Ryu et al. (2026), arXiv:2609.07064, main results (CC BY 4.0).<\/figcaption><\/figure>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83d\udd2c What the study shows<\/h2>\n\n\n<p class=\"wp-block-paragraph\">Among Qwen-based models, 3B outperforms SpatialLadder by 2.7% (3.2% in another comparison) with a simpler synthetic-only pipeline; 7B outperforms SpaceR and Spatial-SSRL despite tiny data.<\/p>\n\n\n<ul class=\"wp-block-list\"><li><strong>Visual cues matter:<\/strong> Removing them drops MindCube by 7.9%\/7.3% (direct\/reason) and MMSI-Bench by 1.9%\/4.8%.<\/li><li><strong>Sample efficiency:<\/strong> 15K synthetic problems punch above real-scene annotation cost.<\/li><\/ul>\n\n\n<h2 class=\"wp-block-heading\">\u2705 Why it matters: cheap spatial foundations<\/h2>\n\n\n<p class=\"wp-block-paragraph\">If toy blocks teach transferable 3D skills, labs can skip some noisy real-scene annotation.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83e\uddea Method in brief<\/h2>\n\n\n<p class=\"wp-block-paragraph\">Synthetic block-stacking dataset; train Qwen-scale LVLMs; evaluate MindCube, MMSI-Bench and related spatial suites.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\u2753 What remains open<\/h2>\n\n<ul class=\"wp-block-list\"><li><strong>Synthetic gap:<\/strong> Real scenes remain harder.<\/li><li><strong>Teaser human 95%:<\/strong> Single illustrative item, not a full human baseline.<\/li><\/ul>\n\n\n<h2 class=\"wp-block-heading\">\ud83e\udded Context<\/h2>\n\n\n<p class=\"wp-block-paragraph\">Related spatial VLM work in <a href=\"\/news\/category\/studies\/multimodal\">Multimodal<\/a>.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83d\udee0\ufe0f What it means for multimodal trainers<\/h2>\n\n<ul class=\"wp-block-list\"><li><strong>Pretrain spatial skills synthetically:<\/strong> Then adapt.<\/li><li><strong>Keep visual cues ablating:<\/strong> They still help.<\/li><li><strong>Compare sample efficiency:<\/strong> 15K can beat heavier pipelines.<\/li><\/ul>\n\n\n<h2 class=\"wp-block-heading\">\u270d\ufe0f About the analysis<\/h2>\n\n\n<p class=\"wp-block-paragraph\">This article is based on the full text of the preprint &#8220;SpatialBlock: Enhancing Spatial Intelligence in LVLMs via Synthetic Block-Stacking Problem&#8221; (arXiv:2609.07064v1, submitted September 7, 2026). All numbers are taken from the paper&#8217;s text, tables and figures; i10X did not rerun the experiments. The chart was redrawn by i10X from main results of the paper, which is published under a CC BY 4.0 license.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83d\udd2d i10x Perspective<\/h2>\n\n\n<p class=\"wp-block-paragraph\">Children learn space with blocks long before street scenes. SpatialBlock suggests LVLMs might too \u2014 and that 15,000 carefully designed toys can out-teach noisier real geometry labels.<\/p>\n\n","protected":false},"excerpt":{"rendered":"<p>SpatialBlock-15k: Qwen-based 4B model reaches 51.3% (+25.1 over backbone). 7B gains 17.6% over backbone; 3B beats SpatialLadder by 2.7\u20133.2% using only synthetic blocks.<\/p>\n","protected":false},"author":2,"featured_media":6324,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"paper_title":"SpatialBlock: Enhancing Spatial Intelligence in LVLMs via Synthetic Block-Stacking Problem","paper_url":"https:\/\/arxiv.org\/abs\/2609.07064","pdf_url":"https:\/\/arxiv.org\/pdf\/2609.07064v1","doi":"","arxiv_id":"2609.07064","authors":"Soohyun Ryu, Sohee Kim, Eunho Yang","institutions":"KAIST; AITRICS","paper_date":"2026-09-07","venue":"arXiv (cs.CV), version 1","peer_reviewed":false,"code_url":"","data_url":"","funding_coi":"KAIST & AITRICS.","figure_license":"Chart drawn by i10X from main results (CC BY 4.0)","paper_license":"CC BY 4.0","replication":"no","study_models":"Qwen-based 3B\/4B\/7B LVLMs vs SpatialLadder, SpaceR, Spatial-SSRL","study_sample":"SpatialBlock-15k plus MindCube\/MMSI-Bench transfers","footnotes":"","_yoast_wpseo_primary_category":""},"categories":[58,52],"tags":[],"companies":[37,38],"products":[49],"class_list":["post-6174","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-multimodal","category-studies","company-google","company-open-ai","product-gemini"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v28.7 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>SpatialBlock: Synthetic Blocks for LVLM Spatial Intelligence - i10X News<\/title>\n<meta name=\"description\" content=\"SpatialBlock-15k teaches LVLMs 3D spatial skills. 4B model hits 51.3%; beats heavier baselines with 15K synthetic problems.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"SpatialBlock: Synthetic Blocks for LVLM Spatial Intelligence - i10X News\" \/>\n<meta property=\"og:description\" content=\"SpatialBlock-15k teaches LVLMs 3D spatial skills. 4B model hits 51.3%; beats heavier baselines with 15K synthetic problems.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence\" \/>\n<meta property=\"og:site_name\" content=\"i10X News\" \/>\n<meta property=\"article:published_time\" content=\"2026-10-19T16:49:47+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-10-19T17:25:14+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-title.webp\" \/>\n\t<meta property=\"og:image:width\" content=\"1344\" \/>\n\t<meta property=\"og:image:height\" content=\"768\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/webp\" \/>\n<meta name=\"author\" content=\"Christopher Ort\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@i10x_ai\" \/>\n<meta name=\"twitter:site\" content=\"@i10x_ai\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Christopher Ort\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"2 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"NewsArticle\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence\"},\"author\":{\"name\":\"Christopher Ort\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#\\\/schema\\\/person\\\/d410bb7ee72febefca2c9c369939219a\"},\"headline\":\"Teaching LVLMs Spatial Skills With 15K Toy Blocks Beats Heavier Real-Scene Pipelines\",\"datePublished\":\"2026-10-19T16:49:47+00:00\",\"dateModified\":\"2026-10-19T17:25:14+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence\"},\"wordCount\":361,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/spatialblock-synthetic-block-stacking-spatial-intelligence-title.webp\",\"articleSection\":[\"Multimodal\",\"Studies\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence\",\"name\":\"SpatialBlock: Synthetic Blocks for LVLM Spatial Intelligence - i10X News\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/spatialblock-synthetic-block-stacking-spatial-intelligence-title.webp\",\"datePublished\":\"2026-10-19T16:49:47+00:00\",\"dateModified\":\"2026-10-19T17:25:14+00:00\",\"description\":\"SpatialBlock-15k teaches LVLMs 3D spatial skills. 4B model hits 51.3%; beats heavier baselines with 15K synthetic problems.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence#primaryimage\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/spatialblock-synthetic-block-stacking-spatial-intelligence-title.webp\",\"contentUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/spatialblock-synthetic-block-stacking-spatial-intelligence-title.webp\",\"width\":1344,\"height\":768,\"caption\":\"Illustration: i10X\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/spatialblock-synthetic-block-stacking-spatial-intelligence#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/news.i10x.ai\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Teaching LVLMs Spatial Skills With 15K Toy Blocks Beats Heavier Real-Scene Pipelines\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#website\",\"url\":\"https:\\\/\\\/news.i10x.ai\\\/\",\"name\":\"i10X News\",\"description\":\"Stay updated with the latest news and announcements from i10X.\",\"publisher\":{\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/news.i10x.ai\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#organization\",\"name\":\"i10X\",\"url\":\"https:\\\/\\\/news.i10x.ai\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/i10x-logo-180.png\",\"contentUrl\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/i10x-logo-180.png\",\"width\":180,\"height\":180,\"caption\":\"i10X\"},\"image\":{\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/x.com\\\/i10x_ai\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/i10x-ai\",\"https:\\\/\\\/instagram.com\\\/i10x.ai\",\"https:\\\/\\\/www.youtube.com\\\/@i10X_ai_official\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#\\\/schema\\\/person\\\/d410bb7ee72febefca2c9c369939219a\",\"name\":\"Christopher Ort\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"contentUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"caption\":\"Christopher Ort\"},\"description\":\"Christopher Ort handles SEO, AI agent workflows and PR & partnerships at i10x.ai, the all-in-one AI workspace built by i10X Global Pte. Ltd. in Singapore. He is a graduate of Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU). He is the author of \u201cThe Style Penalty\u201d, a study published on SSRN that analyses 1,576 evaluation decisions by four frontier AI models to show how AI-written CVs fare in automated hiring screens; the research has been covered by heise online and t3n.\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/author\\\/christopher-ort\",\"sameAs\":[\"https:\\\/\\\/i10x.ai\\\/blog\\\/author\\\/christopher-ort\"],\"worksFor\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#organization\"},\"alumniOf\":{\"@type\":\"EducationalOrganization\",\"name\":\"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU)\",\"url\":\"https:\\\/\\\/www.fau.eu\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"SpatialBlock: Synthetic Blocks for LVLM Spatial Intelligence - i10X News","description":"SpatialBlock-15k teaches LVLMs 3D spatial skills. 4B model hits 51.3%; beats heavier baselines with 15K synthetic problems.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence","og_locale":"en_US","og_type":"article","og_title":"SpatialBlock: Synthetic Blocks for LVLM Spatial Intelligence - i10X News","og_description":"SpatialBlock-15k teaches LVLMs 3D spatial skills. 4B model hits 51.3%; beats heavier baselines with 15K synthetic problems.","og_url":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence","og_site_name":"i10X News","article_published_time":"2026-10-19T16:49:47+00:00","article_modified_time":"2026-10-19T17:25:14+00:00","og_image":[{"width":1344,"height":768,"url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-title.webp","type":"image\/webp"}],"author":"Christopher Ort","twitter_card":"summary_large_image","twitter_creator":"@i10x_ai","twitter_site":"@i10x_ai","twitter_misc":{"Written by":"Christopher Ort","Est. reading time":"2 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"NewsArticle","@id":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence#article","isPartOf":{"@id":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence"},"author":{"name":"Christopher Ort","@id":"https:\/\/news.i10x.ai\/#\/schema\/person\/d410bb7ee72febefca2c9c369939219a"},"headline":"Teaching LVLMs Spatial Skills With 15K Toy Blocks Beats Heavier Real-Scene Pipelines","datePublished":"2026-10-19T16:49:47+00:00","dateModified":"2026-10-19T17:25:14+00:00","mainEntityOfPage":{"@id":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence"},"wordCount":361,"commentCount":0,"publisher":{"@id":"https:\/\/news.i10x.ai\/#organization"},"image":{"@id":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence#primaryimage"},"thumbnailUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-title.webp","articleSection":["Multimodal","Studies"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence#respond"]}]},{"@type":"WebPage","@id":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence","url":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence","name":"SpatialBlock: Synthetic Blocks for LVLM Spatial Intelligence - i10X News","isPartOf":{"@id":"https:\/\/news.i10x.ai\/#website"},"primaryImageOfPage":{"@id":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence#primaryimage"},"image":{"@id":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence#primaryimage"},"thumbnailUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-title.webp","datePublished":"2026-10-19T16:49:47+00:00","dateModified":"2026-10-19T17:25:14+00:00","description":"SpatialBlock-15k teaches LVLMs 3D spatial skills. 4B model hits 51.3%; beats heavier baselines with 15K synthetic problems.","breadcrumb":{"@id":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence#primaryimage","url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-title.webp","contentUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/spatialblock-synthetic-block-stacking-spatial-intelligence-title.webp","width":1344,"height":768,"caption":"Illustration: i10X"},{"@type":"BreadcrumbList","@id":"https:\/\/i10x.ai\/news\/spatialblock-synthetic-block-stacking-spatial-intelligence#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/news.i10x.ai\/"},{"@type":"ListItem","position":2,"name":"Teaching LVLMs Spatial Skills With 15K Toy Blocks Beats Heavier Real-Scene Pipelines"}]},{"@type":"WebSite","@id":"https:\/\/news.i10x.ai\/#website","url":"https:\/\/news.i10x.ai\/","name":"i10X News","description":"Stay updated with the latest news and announcements from i10X.","publisher":{"@id":"https:\/\/news.i10x.ai\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/news.i10x.ai\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/news.i10x.ai\/#organization","name":"i10X","url":"https:\/\/news.i10x.ai\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/news.i10x.ai\/#\/schema\/logo\/image\/","url":"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/10\/i10x-logo-180.png","contentUrl":"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/10\/i10x-logo-180.png","width":180,"height":180,"caption":"i10X"},"image":{"@id":"https:\/\/news.i10x.ai\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/x.com\/i10x_ai","https:\/\/www.linkedin.com\/company\/i10x-ai","https:\/\/instagram.com\/i10x.ai","https:\/\/www.youtube.com\/@i10X_ai_official"]},{"@type":"Person","@id":"https:\/\/news.i10x.ai\/#\/schema\/person\/d410bb7ee72febefca2c9c369939219a","name":"Christopher Ort","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","contentUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","caption":"Christopher Ort"},"description":"Christopher Ort handles SEO, AI agent workflows and PR & partnerships at i10x.ai, the all-in-one AI workspace built by i10X Global Pte. Ltd. in Singapore. He is a graduate of Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU). He is the author of \u201cThe Style Penalty\u201d, a study published on SSRN that analyses 1,576 evaluation decisions by four frontier AI models to show how AI-written CVs fare in automated hiring screens; the research has been covered by heise online and t3n.","url":"https:\/\/i10x.ai\/news\/author\/christopher-ort","sameAs":["https:\/\/i10x.ai\/blog\/author\/christopher-ort"],"worksFor":{"@id":"https:\/\/i10x.ai\/news\/#organization"},"alumniOf":{"@type":"EducationalOrganization","name":"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU)","url":"https:\/\/www.fau.eu"}}]}},"_links":{"self":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/6174","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/comments?post=6174"}],"version-history":[{"count":2,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/6174\/revisions"}],"predecessor-version":[{"id":6325,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/6174\/revisions\/6325"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/media\/6324"}],"wp:attachment":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/media?parent=6174"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/categories?post=6174"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/tags?post=6174"},{"taxonomy":"company","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/companies?post=6174"},{"taxonomy":"product","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/products?post=6174"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}