{"id":6156,"date":"2026-10-19T16:49:33","date_gmt":"2026-10-19T16:49:33","guid":{"rendered":"https:\/\/i10xnews.kinsta.cloud\/?p=6156"},"modified":"2026-10-19T17:22:43","modified_gmt":"2026-10-19T17:22:43","slug":"diffusion-augmented-llms-lossless-speedups","status":"publish","type":"post","link":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups","title":{"rendered":"Diffusion-Augmented LLMs Deliver Up to 3\u00d7 Lossless Speedups Over Autoregressive Decoding"},"content":{"rendered":"\n<h2 class=\"wp-block-heading\">\u26a1 Quick Take<\/h2>\n\n<ul class=\"wp-block-list\"><li><strong>The finding:<\/strong> Diffusion-augmented LLMs keep an AR distribution but draw multiple tokens in parallel via lightweight diffusion weights trained by Diffusion Distillation.<\/li><li><strong>Throughput:<\/strong> Uno exceeds leading speculative-decoding methods at every batch size and reaches up to 3\u00d7 speedups over the base AR model, including at max device batch.<\/li><li><strong>Training too:<\/strong> Up to 2\u00d7 speedup at the largest AR-supported batch size for inference and end-to-end RL training.<\/li><\/ul>\n\n\n<p class=\"wp-block-paragraph\">Autoregressive decoding is sequential and slow. Researchers from the Institute of Foundation Models, UIUC, Cornell Tech, Harvard and Cerebras introduce models that still define an AR distribution \u2014 so quality targets stay the same \u2014 while using discrete diffusion to emit several tokens at once.<\/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\/diffusion-augmented-llms-lossless-speedups-speed.webp\" alt=\"Stat card: up to 3 times speedup versus AR, beats speculative decoding at all batch sizes, up to 2 times at max batch including RL, lightweight diffusion weights.\" class=\"wp-image-6155\" srcset=\"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/diffusion-augmented-llms-lossless-speedups-speed.webp 1344w, https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/diffusion-augmented-llms-lossless-speedups-speed-300x171.webp 300w, https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/diffusion-augmented-llms-lossless-speedups-speed-1024x585.webp 1024w, https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/diffusion-augmented-llms-lossless-speedups-speed-768x439.webp 768w\" sizes=\"auto, (max-width: 1344px) 100vw, 1344px\" \/><figcaption class=\"wp-element-caption\">i10X chart based on data from Sahoo et al. (2026), arXiv:2609.04010, abstract and Sec 5.1 (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\">AR weights train with standard next-token prediction; diffusion weights are a light add-on. \u03a8-Spec samplers enable lossless acceleration and test-time scaling at fixed context without a separate draft model. Pretraining recipes span 21.9T tokens at 8K up through staged extensions to 512K.<\/p>\n\n\n<ul class=\"wp-block-list\"><li><strong>Negligible distillation overhead:<\/strong> Diffusion Distillation slots into existing pipelines.<\/li><li><strong>No separate drafter:<\/strong> Unlike speculative decoding&#8217;s draft+verify stack.<\/li><\/ul>\n\n\n<h2 class=\"wp-block-heading\">\u2705 Why it matters: lossless is the bar<\/h2>\n\n\n<p class=\"wp-block-paragraph\">Many accelerators trade quality. Keeping the AR distribution makes &#8220;faster&#8221; mean the same model, not a cheaper approximation.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83e\uddea Method in brief<\/h2>\n\n\n<p class=\"wp-block-paragraph\">Decoupled AR and diffusion parameters; \u03a8-Spec sampling family; throughput vs speculative decoding across batch sizes.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\u2753 What remains open<\/h2>\n\n<ul class=\"wp-block-list\"><li><strong>Hardware\/batch specifics:<\/strong> Peak speedups depend on device limits.<\/li><li><strong>Implementation maturity:<\/strong> New class of models \u2014 tooling still catching up.<\/li><\/ul>\n\n\n<h2 class=\"wp-block-heading\">\ud83e\udded Context<\/h2>\n\n\n<p class=\"wp-block-paragraph\">Related decoding acceleration in <a href=\"\/news\/category\/studies\/efficiency-training\">Efficiency &amp; Training<\/a>.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83d\udee0\ufe0f What it means for serving and RL<\/h2>\n\n<ul class=\"wp-block-list\"><li><strong>Consider diffusion heads:<\/strong> Especially when speculative decoding saturates.<\/li><li><strong>Accelerate RL rollouts:<\/strong> 2\u00d7 at max batch is training-relevant.<\/li><li><strong>Demand lossless checks:<\/strong> Match AR distribution, not only BLEU.<\/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;Unlocking Lossless Speedups in LLMs via Discrete Diffusion&#8221; (arXiv:2609.04010v1, submitted September 3, 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 abstract and Sec 5.1 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\">Speculative decoding made parallelism mainstream by verifying drafts. Diffusion-augmented LLMs ask whether the draft should live inside the same model. If the 3\u00d7 claims generalize, the sequential tax on LLM decoding finally has a serious in-distribution rival.<\/p>\n\n","protected":false},"excerpt":{"rendered":"<p>Uno adds lightweight diffusion weights to an AR LLM for parallel multi-token draws. It beats speculative decoding throughput at every batch size and speeds RL training up to 2\u00d7 at max batch.<\/p>\n","protected":false},"author":2,"featured_media":6306,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"paper_title":"Unlocking Lossless Speedups in LLMs via Discrete Diffusion","paper_url":"https:\/\/arxiv.org\/abs\/2609.04010","pdf_url":"https:\/\/arxiv.org\/pdf\/2609.04010v1","doi":"","arxiv_id":"2609.04010","authors":"Subham Sekhar Sahoo, Lingjie Chen, Khiem Pham, Jonathan Geuter, Chaitanya Dwivedi, Varad Pimpalkhute, Yash Akhauri, Alexander Moreno, Mikhail Yurochkin, Zhenting Wang, Mostafa Elhoushi, Nolan Dey et al. (17 authors)","institutions":"Institute of Foundation Models; UIUC; Cornell Tech; Harvard; Cerebras Systems","paper_date":"2026-09-03","venue":"arXiv (cs.LG), version 1","peer_reviewed":false,"code_url":"","data_url":"","funding_coi":"Multi-institution; Cerebras co-authors.","figure_license":"Chart drawn by i10X from abstract and Sec 5.1 (CC BY 4.0)","paper_license":"CC BY 4.0","replication":"no","study_models":"Uno diffusion-augmented LLMs vs AR and speculative decoding baselines","study_sample":"Throughput across batch sizes; RL training acceleration tests","footnotes":"","_yoast_wpseo_primary_category":""},"categories":[57,52],"tags":[],"companies":[],"products":[],"class_list":["post-6156","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-efficiency-training","category-studies"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v28.7 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>Lossless LLM Speedups via Discrete Diffusion (Uno) - i10X News<\/title>\n<meta name=\"description\" content=\"Diffusion-augmented LLMs: up to 3\u00d7 speedups vs AR, beating speculative decoding at all batch sizes. \u03a8-Spec samplers keep quality.\" \/>\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\/diffusion-augmented-llms-lossless-speedups\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Lossless LLM Speedups via Discrete Diffusion (Uno) - i10X News\" \/>\n<meta property=\"og:description\" content=\"Diffusion-augmented LLMs: up to 3\u00d7 speedups vs AR, beating speculative decoding at all batch sizes. \u03a8-Spec samplers keep quality.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups\" \/>\n<meta property=\"og:site_name\" content=\"i10X News\" \/>\n<meta property=\"article:published_time\" content=\"2026-10-19T16:49:33+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-10-19T17:22:43+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/diffusion-augmented-llms-lossless-speedups-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\\\/diffusion-augmented-llms-lossless-speedups#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups\"},\"author\":{\"name\":\"Christopher Ort\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#\\\/schema\\\/person\\\/d410bb7ee72febefca2c9c369939219a\"},\"headline\":\"Diffusion-Augmented LLMs Deliver Up to 3\u00d7 Lossless Speedups Over Autoregressive Decoding\",\"datePublished\":\"2026-10-19T16:49:33+00:00\",\"dateModified\":\"2026-10-19T17:22:43+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups\"},\"wordCount\":411,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/diffusion-augmented-llms-lossless-speedups-title.webp\",\"articleSection\":[\"Efficiency &amp; Training\",\"Studies\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups\",\"name\":\"Lossless LLM Speedups via Discrete Diffusion (Uno) - i10X News\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/diffusion-augmented-llms-lossless-speedups-title.webp\",\"datePublished\":\"2026-10-19T16:49:33+00:00\",\"dateModified\":\"2026-10-19T17:22:43+00:00\",\"description\":\"Diffusion-augmented LLMs: up to 3\u00d7 speedups vs AR, beating speculative decoding at all batch sizes. \u03a8-Spec samplers keep quality.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups#primaryimage\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/diffusion-augmented-llms-lossless-speedups-title.webp\",\"contentUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/diffusion-augmented-llms-lossless-speedups-title.webp\",\"width\":1344,\"height\":768,\"caption\":\"Illustration: i10X\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/diffusion-augmented-llms-lossless-speedups#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/news.i10x.ai\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Diffusion-Augmented LLMs Deliver Up to 3\u00d7 Lossless Speedups Over Autoregressive Decoding\"}]},{\"@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":"Lossless LLM Speedups via Discrete Diffusion (Uno) - i10X News","description":"Diffusion-augmented LLMs: up to 3\u00d7 speedups vs AR, beating speculative decoding at all batch sizes. \u03a8-Spec samplers keep quality.","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\/diffusion-augmented-llms-lossless-speedups","og_locale":"en_US","og_type":"article","og_title":"Lossless LLM Speedups via Discrete Diffusion (Uno) - i10X News","og_description":"Diffusion-augmented LLMs: up to 3\u00d7 speedups vs AR, beating speculative decoding at all batch sizes. \u03a8-Spec samplers keep quality.","og_url":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups","og_site_name":"i10X News","article_published_time":"2026-10-19T16:49:33+00:00","article_modified_time":"2026-10-19T17:22:43+00:00","og_image":[{"width":1344,"height":768,"url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/diffusion-augmented-llms-lossless-speedups-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\/diffusion-augmented-llms-lossless-speedups#article","isPartOf":{"@id":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups"},"author":{"name":"Christopher Ort","@id":"https:\/\/news.i10x.ai\/#\/schema\/person\/d410bb7ee72febefca2c9c369939219a"},"headline":"Diffusion-Augmented LLMs Deliver Up to 3\u00d7 Lossless Speedups Over Autoregressive Decoding","datePublished":"2026-10-19T16:49:33+00:00","dateModified":"2026-10-19T17:22:43+00:00","mainEntityOfPage":{"@id":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups"},"wordCount":411,"commentCount":0,"publisher":{"@id":"https:\/\/news.i10x.ai\/#organization"},"image":{"@id":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups#primaryimage"},"thumbnailUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/diffusion-augmented-llms-lossless-speedups-title.webp","articleSection":["Efficiency &amp; Training","Studies"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups#respond"]}]},{"@type":"WebPage","@id":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups","url":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups","name":"Lossless LLM Speedups via Discrete Diffusion (Uno) - i10X News","isPartOf":{"@id":"https:\/\/news.i10x.ai\/#website"},"primaryImageOfPage":{"@id":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups#primaryimage"},"image":{"@id":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups#primaryimage"},"thumbnailUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/diffusion-augmented-llms-lossless-speedups-title.webp","datePublished":"2026-10-19T16:49:33+00:00","dateModified":"2026-10-19T17:22:43+00:00","description":"Diffusion-augmented LLMs: up to 3\u00d7 speedups vs AR, beating speculative decoding at all batch sizes. \u03a8-Spec samplers keep quality.","breadcrumb":{"@id":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups#primaryimage","url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/diffusion-augmented-llms-lossless-speedups-title.webp","contentUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/diffusion-augmented-llms-lossless-speedups-title.webp","width":1344,"height":768,"caption":"Illustration: i10X"},{"@type":"BreadcrumbList","@id":"https:\/\/i10x.ai\/news\/diffusion-augmented-llms-lossless-speedups#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/news.i10x.ai\/"},{"@type":"ListItem","position":2,"name":"Diffusion-Augmented LLMs Deliver Up to 3\u00d7 Lossless Speedups Over Autoregressive Decoding"}]},{"@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\/6156","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=6156"}],"version-history":[{"count":2,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/6156\/revisions"}],"predecessor-version":[{"id":6307,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/6156\/revisions\/6307"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/media\/6306"}],"wp:attachment":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/media?parent=6156"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/categories?post=6156"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/tags?post=6156"},{"taxonomy":"company","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/companies?post=6156"},{"taxonomy":"product","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/products?post=6156"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}