{"id":13513,"date":"2025-11-13T17:39:35","date_gmt":"2025-11-13T17:39:35","guid":{"rendered":"https:\/\/www.yiaho.com\/trm-what-are-tiny-recursive-models-in-ai\/"},"modified":"2025-11-13T17:39:35","modified_gmt":"2025-11-13T17:39:35","slug":"trm-what-are-tiny-recursive-models-in-ai","status":"publish","type":"post","link":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/","title":{"rendered":"TRM: What are Tiny Recursive Models in AI?"},"content":{"rendered":"<p>For the past ten years, artificial intelligence has been dominated by <a href=\"https:\/\/www.yiaho.com\/en\/what-is-a-large-language-model-llm-in-ai\/\" target=\"_blank\" rel=\"noopener\">LLMs<\/a>. GPT-5, Gemini, Claude: models with hundreds of billions of parameters, trained on oceans of data, capable of writing a novel or coding an application in seconds. <\/p>\n<p>However, a Canadian researcher has just demonstrated that comparable, or even superior, performance can be achieved on reasoning tasks with a model 200 times smaller and 200 times more energy-efficient. At Yiaho, we closely follow these advancements and are always ready to innovate. Today, we&#8217;re going to explain the world of Tiny Recursive Models, or TRMs!  <\/p>\n<h2>What is a Tiny Recursive Model?<\/h2>\n<p>A TRM is an extremely compact artificial neural network, with only 1 to 10 million parameters, designed specifically for structured reasoning. Unlike LLMs (Large Language Models) which learn to predict the next word in a sentence, TRMs learn to solve problems step-by-step, like a human solving an equation or a logic puzzle. <\/p>\n<p>The secret? Four technical pillars: <\/p>\n<ul>\n<li><strong>No massive pre-training<\/strong>: gone are the terabytes of text scraped from the internet. The model is trained on a targeted, synthetic dataset, often generated by the model itself. <\/li>\n<li><strong>Deep supervision<\/strong>: each layer of the network receives a direct learning signal, not just the final output. This forces the model to learn useful representations at all levels. <\/li>\n<li><strong>Deep recursion<\/strong>: the network calls itself iteratively, like a function that recalls itself until convergence.<\/li>\n<li><strong>Recursive data augmentation<\/strong>: with each iteration, the model generates new examples from its own errors, enriching its own training corpus.<\/li>\n<\/ul>\n<p><strong>Result<\/strong>: a model that doesn&#8217;t memorize, but truly reasons. For example, the TinyReasoner-7M, with only 7 million parameters, excels at complex puzzles like Sudoku, mazes, or ARC-AGI evaluation, benchmarks where even giants like <a href=\"https:\/\/www.yiaho.com\/en\/gemini-2-5-pro-google-pushes-the-limits-of-ai-reasoning\/\" target=\"_blank\" rel=\"noopener\">Gemini 2.5 Pro<\/a> struggle. <\/p>\n<h2>Backtracking: The Big Difference from LLMs<\/h2>\n<p>Imagine you&#8217;re playing chess. An LLM plays move by move, never going back. If it makes a mistake on the 5th move, the whole game is lost. A TRM, however, can say: &#8220;<em>This path leads to failure. I&#8217;ll go back three moves and try something else.<\/em>&#8221;   <\/p>\n<p><strong>This is recursive backtracking. The model: <\/strong><\/p>\n<ul>\n<li>Generates a partial reasoning sequence.<\/li>\n<li>Evaluates it with an internal scoring function (e.g., &#8220;Does this step follow logical rules?&#8221;).<\/li>\n<li>If the score is too low, it goes back, modifies a previous step, and restarts.<\/li>\n<li>It repeats until a valid solution is obtained.<\/li>\n<\/ul>\n<p><strong>Major consequence<\/strong>: the TRM does not generate its response <a href=\"https:\/\/www.yiaho.com\/en\/what-is-an-ai-token-definition-and-explanation\/\" target=\"_blank\" rel=\"noopener\">token by token<\/a>. It works internally, refines, corrects, optimizes&#8230; then delivers a complete, coherent, and final answer in one block. It&#8217;s like a mathematician filling pages of rough work before showing you the final, unblemished proof.  <\/p>\n<p>This approach drastically reduces &#8220;<a href=\"https:\/\/www.yiaho.com\/en\/ai-hallucination-why-does-chatgpt-sometimes-invent-answers\/\" target=\"_blank\" rel=\"noopener\">hallucinations<\/a>,&#8221; those absurd errors that LLMs often make, and allows for much higher precision on tasks requiring pure logic.<\/p>\n<h2>The Paper That Shook the Markets<\/h2>\n<p>It all began on October 6, 2025, with the publication of &#8220;<em>Less is More: Recursive Reasoning with Tiny Networks<\/em>&#8221; by Alexia Jolicoeur-Martineau, principal researcher at Samsung SAIT AI Lab in Montreal. (Paper available in the sources at the bottom of the page). <\/p>\n<p>In this document, she presents the Tiny Recursive Model (TRM), inspired by the Hierarchical Reasoning Model (HRM), but simplified to the extreme. The 7-million-parameter TRM: <\/p>\n<ul>\n<li>Achieves 45% on ARC-AGI-1 and 8% on ARC-AGI-2, outperforming models like <a href=\"https:\/\/www.yiaho.com\/en\/deepseek-the-chinese-ai-that-could-outperform-chatgpt\/\" target=\"_blank\" rel=\"noopener\">DeepSeek-R1<\/a> or Gemini 2.5 Pro.<\/li>\n<li>Beats LLMs on tasks like Sudoku, mazes, and abstract reasoning, with only 1000 training examples.<\/li>\n<li>Runs on a single <a href=\"https:\/\/www.yiaho.com\/cest-quoi-un-gpu-en-ia-definition-et-explication\/\" target=\"_blank\" rel=\"noopener\">GPU<\/a> RTX 4090 for inference, and trains in a few hours.<\/li>\n<li>Consumes 0.5 watt-hours per inference, compared to 100+ for an equivalent LLM.<\/li>\n<\/ul>\n<p>Two days after publication, Samsung&#8217;s stock jumped 10.3% on the Seoul Stock Exchange! Analysts are already talking about revolutionary <a href=\"https:\/\/www.yiaho.com\/en\/edge-ai-artificial-intelligence-thats-showing-up-everywhere-and-why-it-changes-everything\/\" target=\"_blank\" rel=\"noopener\">edge AI<\/a>: a model capable of reasoning like a human, but fitting into a smartwatch or an industrial sensor. The buzz spread on Reddit and Medium, where open-source AI communities are enthusiastic about this proof that &#8220;less is more.&#8221; In plain English, we can say that simplicity can often be more effective than complexity!   <\/p>\n<h2>Why TRMs Can Change Everything?<\/h2>\n<h3>1. They work without the cloud<\/h3>\n<p>No need to send your data to a remote server. The model runs locally, on your phone, your car, your fridge. Privacy guaranteed, and near-zero latency.  <\/p>\n<h3>2. They are energy efficient<\/h3>\n<p>A TRM consumes the equivalent of an LED bulb. An LLM? The equivalent of a small building. In a world where <a href=\"https:\/\/www.yiaho.com\/en\/does-ai-pollute-the-answer-is-surprising\/\" target=\"_blank\" rel=\"noopener\">AI already accounts for 3% of global electricity consumption<\/a>, this is crucial for mass deployment.   <\/p>\n<h3>3. They are reliable by design<\/h3>\n<p>Thanks to backtracking, the model almost never hallucinates on structured tasks. It checks, corrects, validates. Ideal for medicine (logical diagnosis), finance (calculation verification), security (route planning), or even video games.  <\/p>\n<h3>4. They cost almost nothing to train<\/h3>\n<p>A few thousand dollars and a weekend are enough. Any lab, startup, or student can create one. The code is already open-source on GitHub, under the <em>SamsungSAILMontreal\/TinyRecursiveModels<\/em> repo, with a ready-to-use PyTorch implementation!  <\/p>\n<p>In comparison, training an LLM like GPT-3 costs millions and requires GPU farms. TRMs democratize reasoning AI, making what was once reserved for tech giants accessible. <\/p>\n<h2>TRMs: The Future of AI?<\/h2>\n<p>TRMs cannot write poems, hold fluid conversations, or generate creative images. Their domain? Pure reasoning: math, logic, planning, diagnosis, games, code verification. They shine where LLMs stumble, such as on ARC-AGI&#8217;s abstract puzzles, which test general intelligence without massive data bias.   <\/p>\n<p><strong>But the future is promising:<\/strong><\/p>\n<ul>\n<li>TRM + LLM Hybrids: the TRM handles the heavy reasoning, the LLM drafts the response in natural language for a more human interaction.<\/li>\n<li>Integration into Galaxy S26: Samsung is already preparing embedded assistants based on TRMs for local and energy-efficient AI.<\/li>\n<li>Expanding Open-Source: the paper&#8217;s code is on arXiv, and GitHub forks are proliferating. Tutorials for training your own TRM in 3 hours are already emerging on Medium. <\/li>\n<\/ul>\n<p>With the recent excitement, from heated discussions on X to analyses on Hugging Face, TRMs could well mark the turning point towards more efficient and accessible AI.<\/p>\n<h2>In summary: less really is more!<\/h2>\n<p>Tiny Recursive Models are not a passing fad. They are proof that in AI, size is not the solution to everything. With a few million parameters, good architecture, and a lot of ingenuity, complex problems can be solved reliably, quickly, and economically.  <\/p>\n<p>Alexia Jolicoeur-Martineau has demonstrated it: recursion and backtracking transform &#8220;small&#8221; models into reasoning champions. TRMs are AI that thinks before it speaks. Which reminds us of a forgotten lesson: sometimes, intelligence is born from constraint.  <\/p>\n<p>Source: <a href=\"https:\/\/arxiv.org\/pdf\/2510.04871\" target=\"_blank\" rel=\"noopener\">Arxiv \/ Less is More: Recursive Reasoning with Tiny Networks<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>For the past ten years, artificial intelligence has been dominated by LLMs. GPT-5, Gemini, Claude: models with hundreds of billions of parameters, trained on oceans of data, capable of writing a novel or coding an application in seconds. However, a Canadian researcher has just demonstrated that comparable, or even superior, performance can be achieved on&hellip;&nbsp;<a href=\"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/\" rel=\"bookmark\">Read More &raquo;<span class=\"screen-reader-text\">TRM: What are Tiny Recursive Models in AI?<\/span><\/a><\/p>\n","protected":false},"author":4,"featured_media":13515,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"neve_meta_sidebar":"","neve_meta_container":"","neve_meta_enable_content_width":"off","neve_meta_content_width":70,"neve_meta_title_alignment":"","neve_meta_author_avatar":"","neve_post_elements_order":"","neve_meta_disable_header":"","neve_meta_disable_footer":"","neve_meta_disable_title":"","neve_meta_reading_time":"","footnotes":""},"categories":[50],"tags":[],"class_list":["post-13513","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-glossary"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.5 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>TRM: What are Tiny Recursive Models in AI?<\/title>\n<meta name=\"description\" content=\"What are TRMs? Tiny AI models that outperform GPT on abstract reasoning and consume 200 times less energy!\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"TRM: What are Tiny Recursive Models in AI?\" \/>\n<meta property=\"og:description\" content=\"What are TRMs? Tiny AI models that outperform GPT on abstract reasoning and consume 200 times less energy!\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/\" \/>\n<meta property=\"og:site_name\" content=\"YIAHO\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/yiaho.ia.gratuite\" \/>\n<meta property=\"article:published_time\" content=\"2025-11-13T17:39:35+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/def-Tiny-Recursive-Models.webp\" \/>\n\t<meta property=\"og:image:width\" content=\"1200\" \/>\n\t<meta property=\"og:image:height\" content=\"800\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/webp\" \/>\n<meta name=\"author\" content=\"G. de Yiaho\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@Yiaho_AI\" \/>\n<meta name=\"twitter:site\" content=\"@Yiaho_AI\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"G. de Yiaho\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"5 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/\"},\"author\":{\"name\":\"G. de Yiaho\",\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/#\\\/schema\\\/person\\\/09fcc2462849b463e2b2511013897d80\"},\"headline\":\"TRM: What are Tiny Recursive Models in AI?\",\"datePublished\":\"2025-11-13T17:39:35+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/\"},\"wordCount\":1072,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/www.yiaho.com\\\/wp-content\\\/uploads\\\/2025\\\/11\\\/def-Tiny-Recursive-Models.webp\",\"articleSection\":[\"AI Glossary\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/\",\"url\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/\",\"name\":\"TRM: What are Tiny Recursive Models in AI?\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/www.yiaho.com\\\/wp-content\\\/uploads\\\/2025\\\/11\\\/def-Tiny-Recursive-Models.webp\",\"datePublished\":\"2025-11-13T17:39:35+00:00\",\"description\":\"What are TRMs? Tiny AI models that outperform GPT on abstract reasoning and consume 200 times less energy!\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/#primaryimage\",\"url\":\"https:\\\/\\\/www.yiaho.com\\\/wp-content\\\/uploads\\\/2025\\\/11\\\/def-Tiny-Recursive-Models.webp\",\"contentUrl\":\"https:\\\/\\\/www.yiaho.com\\\/wp-content\\\/uploads\\\/2025\\\/11\\\/def-Tiny-Recursive-Models.webp\",\"width\":1200,\"height\":800,\"caption\":\"Tiny Recursive Models (TRMs) are revolutionizing AI: with only 7 million parameters, they reason better than large models. Illustration image. Image credit Yiaho\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/trm-what-are-tiny-recursive-models-in-ai\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Accueil\",\"item\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/home\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"TRM: What are Tiny Recursive Models in AI?\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/#website\",\"url\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/\",\"name\":\"YIAHO\",\"description\":\"L&#039;intelligence Artificielle gratuite en ligne et fran\u00e7aise\",\"publisher\":{\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/#organization\",\"name\":\"Yiaho\",\"url\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/www.yiaho.com\\\/wp-content\\\/uploads\\\/2025\\\/11\\\/YIAHO-logo.webp\",\"contentUrl\":\"https:\\\/\\\/www.yiaho.com\\\/wp-content\\\/uploads\\\/2025\\\/11\\\/YIAHO-logo.webp\",\"width\":417,\"height\":424,\"caption\":\"Yiaho\"},\"image\":{\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/www.facebook.com\\\/yiaho.ia.gratuite\",\"https:\\\/\\\/x.com\\\/Yiaho_AI\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/en\\\/#\\\/schema\\\/person\\\/09fcc2462849b463e2b2511013897d80\",\"name\":\"G. de Yiaho\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/www.yiaho.com\\\/wp-content\\\/uploads\\\/2025\\\/11\\\/YIAHO-logo-96x96.webp\",\"url\":\"https:\\\/\\\/www.yiaho.com\\\/wp-content\\\/uploads\\\/2025\\\/11\\\/YIAHO-logo-96x96.webp\",\"contentUrl\":\"https:\\\/\\\/www.yiaho.com\\\/wp-content\\\/uploads\\\/2025\\\/11\\\/YIAHO-logo-96x96.webp\",\"caption\":\"G. de Yiaho\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"TRM: What are Tiny Recursive Models in AI?","description":"What are TRMs? Tiny AI models that outperform GPT on abstract reasoning and consume 200 times less energy!","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:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/","og_locale":"en_US","og_type":"article","og_title":"TRM: What are Tiny Recursive Models in AI?","og_description":"What are TRMs? Tiny AI models that outperform GPT on abstract reasoning and consume 200 times less energy!","og_url":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/","og_site_name":"YIAHO","article_publisher":"https:\/\/www.facebook.com\/yiaho.ia.gratuite","article_published_time":"2025-11-13T17:39:35+00:00","og_image":[{"width":1200,"height":800,"url":"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/def-Tiny-Recursive-Models.webp","type":"image\/webp"}],"author":"G. de Yiaho","twitter_card":"summary_large_image","twitter_creator":"@Yiaho_AI","twitter_site":"@Yiaho_AI","twitter_misc":{"Written by":"G. de Yiaho","Est. reading time":"5 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/#article","isPartOf":{"@id":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/"},"author":{"name":"G. de Yiaho","@id":"https:\/\/www.yiaho.com\/en\/#\/schema\/person\/09fcc2462849b463e2b2511013897d80"},"headline":"TRM: What are Tiny Recursive Models in AI?","datePublished":"2025-11-13T17:39:35+00:00","mainEntityOfPage":{"@id":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/"},"wordCount":1072,"commentCount":0,"publisher":{"@id":"https:\/\/www.yiaho.com\/en\/#organization"},"image":{"@id":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/#primaryimage"},"thumbnailUrl":"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/def-Tiny-Recursive-Models.webp","articleSection":["AI Glossary"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/","url":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/","name":"TRM: What are Tiny Recursive Models in AI?","isPartOf":{"@id":"https:\/\/www.yiaho.com\/en\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/#primaryimage"},"image":{"@id":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/#primaryimage"},"thumbnailUrl":"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/def-Tiny-Recursive-Models.webp","datePublished":"2025-11-13T17:39:35+00:00","description":"What are TRMs? Tiny AI models that outperform GPT on abstract reasoning and consume 200 times less energy!","breadcrumb":{"@id":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/#primaryimage","url":"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/def-Tiny-Recursive-Models.webp","contentUrl":"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/def-Tiny-Recursive-Models.webp","width":1200,"height":800,"caption":"Tiny Recursive Models (TRMs) are revolutionizing AI: with only 7 million parameters, they reason better than large models. Illustration image. Image credit Yiaho"},{"@type":"BreadcrumbList","@id":"https:\/\/www.yiaho.com\/en\/trm-what-are-tiny-recursive-models-in-ai\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Accueil","item":"https:\/\/www.yiaho.com\/en\/home\/"},{"@type":"ListItem","position":2,"name":"TRM: What are Tiny Recursive Models in AI?"}]},{"@type":"WebSite","@id":"https:\/\/www.yiaho.com\/en\/#website","url":"https:\/\/www.yiaho.com\/en\/","name":"YIAHO","description":"L&#039;intelligence Artificielle gratuite en ligne et fran\u00e7aise","publisher":{"@id":"https:\/\/www.yiaho.com\/en\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.yiaho.com\/en\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/www.yiaho.com\/en\/#organization","name":"Yiaho","url":"https:\/\/www.yiaho.com\/en\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.yiaho.com\/en\/#\/schema\/logo\/image\/","url":"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/YIAHO-logo.webp","contentUrl":"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/YIAHO-logo.webp","width":417,"height":424,"caption":"Yiaho"},"image":{"@id":"https:\/\/www.yiaho.com\/en\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/yiaho.ia.gratuite","https:\/\/x.com\/Yiaho_AI"]},{"@type":"Person","@id":"https:\/\/www.yiaho.com\/en\/#\/schema\/person\/09fcc2462849b463e2b2511013897d80","name":"G. de Yiaho","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/YIAHO-logo-96x96.webp","url":"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/YIAHO-logo-96x96.webp","contentUrl":"https:\/\/www.yiaho.com\/wp-content\/uploads\/2025\/11\/YIAHO-logo-96x96.webp","caption":"G. de Yiaho"}}]}},"_links":{"self":[{"href":"https:\/\/www.yiaho.com\/en\/wp-json\/wp\/v2\/posts\/13513","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.yiaho.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.yiaho.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.yiaho.com\/en\/wp-json\/wp\/v2\/users\/4"}],"replies":[{"embeddable":true,"href":"https:\/\/www.yiaho.com\/en\/wp-json\/wp\/v2\/comments?post=13513"}],"version-history":[{"count":0,"href":"https:\/\/www.yiaho.com\/en\/wp-json\/wp\/v2\/posts\/13513\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.yiaho.com\/en\/wp-json\/wp\/v2\/media\/13515"}],"wp:attachment":[{"href":"https:\/\/www.yiaho.com\/en\/wp-json\/wp\/v2\/media?parent=13513"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.yiaho.com\/en\/wp-json\/wp\/v2\/categories?post=13513"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.yiaho.com\/en\/wp-json\/wp\/v2\/tags?post=13513"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}