{"id":7021,"date":"2025-05-06T18:16:19","date_gmt":"2025-05-06T09:16:19","guid":{"rendered":"https:\/\/aireviewirush.com\/?p=7021"},"modified":"2025-05-06T18:16:19","modified_gmt":"2025-05-06T09:16:19","slug":"one-yr-of-phi-small-language-fashions-making-large-leaps-in-ai","status":"publish","type":"post","link":"https:\/\/aireviewirush.com\/?p=7021","title":{"rendered":"One yr of Phi: Small language fashions making large leaps in AI"},"content":{"rendered":"<p> <br \/>\n<\/p>\n<div>\n<p>\n\t\t\tMicrosoft continues so as to add to the dialog by unveiling its latest fashions, Phi-4-reasoning, Phi-4-reasoning-plus, and Phi-4-mini-reasoning.\u00a0\t\t<\/p>\n<p class=\"wp-block-paragraph\"><strong>A brand new period of AI\u00a0<\/strong><\/p>\n<p class=\"wp-block-paragraph\">One yr in the past, Microsoft launched <strong>small language fashions<\/strong> (SLMs) to prospects with the discharge of <strong>Phi-3<\/strong> on <a href=\"https:\/\/ai.azure.com\/?tid=72f988bf-86f1-41af-91ab-2d7cd011db47\" target=\"_blank\" rel=\"noopener\">Azure AI Foundry<\/a>, leveraging analysis on SLMs to increase the vary of environment friendly AI fashions and instruments obtainable to prospects.\u00a0<\/p>\n<p class=\"wp-block-paragraph\">Right this moment, we&#8217;re excited to introduce <strong>Phi-4-reasoning, Phi-4-reasoning-plus, and Phi-4-mini-reasoning<\/strong>\u2014marking a brand new period for small language fashions and as soon as once more redefining what is feasible with small and environment friendly AI.\u00a0<\/p>\n<aside class=\"cta-block cta-block--align-left cta-block--has-image wp-block-msx-cta\" data-bi-an=\"CTA Block\">\n<div class=\"cta-block__content\">\n<div class=\"cta-block__image-container\">\n\t\t\t\t<img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"683\" src=\"https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/01\/MSC24-ASEAN-developer-Getty-1336501076-rgb-1024x683.jpg\" class=\"cta-block__image\" alt=\"A man sitting at a desk with a computer\" srcset=\"https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/01\/MSC24-ASEAN-developer-Getty-1336501076-rgb-1024x683.jpg 1024w, https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/01\/MSC24-ASEAN-developer-Getty-1336501076-rgb-300x200.jpg 300w, https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/01\/MSC24-ASEAN-developer-Getty-1336501076-rgb-768x512.jpg 768w, https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/01\/MSC24-ASEAN-developer-Getty-1336501076-rgb-1536x1025.jpg 1536w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\"\/>\t\t\t<\/div>\n<div class=\"cta-block__body\">\n<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_53 counter-hierarchy ez-toc-counter ez-toc-grey ez-toc-container-direction\">\n<div class=\"ez-toc-title-container\">\n<p class=\"ez-toc-title \" >Table of Contents<\/p>\n<span class=\"ez-toc-title-toggle\"><a href=\"#\" class=\"ez-toc-pull-right ez-toc-btn ez-toc-btn-xs ez-toc-btn-default ez-toc-toggle\" aria-label=\"Toggle Table of Content\" role=\"button\"><label for=\"item-69e7d9a152c9e\" ><span class=\"\"><span style=\"display:none;\">Toggle<\/span><span class=\"ez-toc-icon-toggle-span\"><svg style=\"fill: #999;color:#999\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" class=\"list-377408\" width=\"20px\" height=\"20px\" viewBox=\"0 0 24 24\" fill=\"none\"><path d=\"M6 6H4v2h2V6zm14 0H8v2h12V6zM4 11h2v2H4v-2zm16 0H8v2h12v-2zM4 16h2v2H4v-2zm16 0H8v2h12v-2z\" fill=\"currentColor\"><\/path><\/svg><svg style=\"fill: #999;color:#999\" class=\"arrow-unsorted-368013\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" width=\"10px\" height=\"10px\" viewBox=\"0 0 24 24\" version=\"1.2\" baseProfile=\"tiny\"><path d=\"M18.2 9.3l-6.2-6.3-6.2 6.3c-.2.2-.3.4-.3.7s.1.5.3.7c.2.2.4.3.7.3h11c.3 0 .5-.1.7-.3.2-.2.3-.5.3-.7s-.1-.5-.3-.7zM5.8 14.7l6.2 6.3 6.2-6.3c.2-.2.3-.5.3-.7s-.1-.5-.3-.7c-.2-.2-.4-.3-.7-.3h-11c-.3 0-.5.1-.7.3-.2.2-.3.5-.3.7s.1.5.3.7z\"\/><\/svg><\/span><\/span><\/label><input aria-label=\"Toggle\" aria-label=\"item-69e7d9a152c9e\"  type=\"checkbox\" id=\"item-69e7d9a152c9e\"><\/a><\/span><\/div>\n<nav><ul class='ez-toc-list ez-toc-list-level-1 ' ><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/aireviewirush.com\/?p=7021\/#Azure_AI_Foundry\" title=\"Azure AI Foundry\">Azure AI Foundry<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/aireviewirush.com\/?p=7021\/#Reasoning_fashions_the_following_step_ahead\" title=\"Reasoning fashions, the following step ahead\">Reasoning fashions, the following step ahead<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/aireviewirush.com\/?p=7021\/#Phi-4-reasoning_and_Phi-4-reasoning-plus\" title=\"Phi-4-reasoning and Phi-4-reasoning-plus\u00a0\">Phi-4-reasoning and Phi-4-reasoning-plus\u00a0<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/aireviewirush.com\/?p=7021\/#Phi-4-mini-reasoning\" title=\"Phi-4-mini-reasoning\">Phi-4-mini-reasoning<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/aireviewirush.com\/?p=7021\/#Security_and_Microsoft%E2%80%99s_strategy_to_accountable_AI\" title=\"Security and Microsoft\u2019s strategy to accountable AI\u00a0\">Security and Microsoft\u2019s strategy to accountable AI\u00a0<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/aireviewirush.com\/?p=7021\/#Be_taught_extra_right_here\" title=\"Be taught extra right here:\u00a0\">Be taught extra right here:\u00a0<\/a><\/li><\/ul><\/nav><\/div>\n<h2 class=\"cta-block__headline\"><span class=\"ez-toc-section\" id=\"Azure_AI_Foundry\"><\/span>Azure AI Foundry<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p class=\"cta-block__text\">Discover the best mannequin for your online business wants, then tinker, tweak, and customise inside a mission to attain all of your AI targets.<\/p>\n<\/p><\/div>\n<\/p><\/div>\n<\/aside>\n<h2 class=\"wp-block-heading\" id=\"reasoning-models-the-next-step-forward\"><span class=\"ez-toc-section\" id=\"Reasoning_fashions_the_following_step_ahead\"><\/span>Reasoning fashions, the following step ahead<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p class=\"wp-block-paragraph\"><strong>Reasoning fashions<\/strong> are skilled to leverage inference-time scaling to carry out advanced duties that demand multi-step decomposition and inner reflection. They excel in mathematical reasoning and are rising because the spine of agentic functions with advanced, multi-faceted duties. Such capabilities are sometimes discovered solely in massive frontier fashions.\u00a0Phi-reasoning fashions introduce a brand new class of small language fashions. Utilizing distillation, reinforcement studying, and high-quality knowledge, these fashions stability dimension and efficiency. They&#8217;re sufficiently small for low-latency environments but preserve robust reasoning capabilities that rival a lot larger fashions. This mix permits even resource-limited units to carry out advanced reasoning duties effectively.<\/p>\n<h2 class=\"wp-block-heading\" id=\"phi-4-reasoning-and-phi-4-reasoning-plus\"><span class=\"ez-toc-section\" id=\"Phi-4-reasoning_and_Phi-4-reasoning-plus\"><\/span>Phi-4-reasoning and Phi-4-reasoning-plus\u00a0<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p class=\"wp-block-paragraph\"><strong>Phi-4-reasoning <\/strong>is a 14-billion parameter open-weight reasoning mannequin that rivals a lot bigger fashions on advanced reasoning duties. Skilled through supervised fine-tuning of Phi-4 on fastidiously curated reasoning demonstrations from OpenAI o3-mini, Phi-4-reasoning generates detailed reasoning chains that successfully leverage extra inference-time compute. The mannequin demonstrates that meticulous knowledge curation and high-quality artificial datasets enable smaller fashions to compete with bigger counterparts.<\/p>\n<p class=\"wp-block-paragraph\"><strong>Phi-4-reasoning-plus<\/strong> builds upon Phi-4-reasoning capabilities, additional skilled with reinforcement studying to make the most of extra inference-time compute, utilizing 1.5x extra tokens than Phi-4-reasoning, to ship larger accuracy.<\/p>\n<p class=\"wp-block-paragraph\">Regardless of their considerably smaller dimension, each fashions obtain higher efficiency than OpenAI o1-mini and DeepSeek-R1-Distill-Llama-70B at most benchmarks, together with mathematical reasoning and Ph.D. stage science questions. They obtain efficiency higher than the total DeepSeek-R1 mannequin (with 671-billion parameters) on the AIME 2025 take a look at, the 2025 qualifier for the USA Math Olympiad. Each fashions can be found on <a href=\"https:\/\/ai.azure.com\/explore\/models\/Phi-4-reasoning\/version\/1\/registry\/azureml?tid=72f988bf-86f1-41af-91ab-2d7cd011db47\" target=\"_blank\" rel=\"noopener\">Azure AI Foundry<\/a> and HuggingFace, <a href=\"https:\/\/huggingface.co\/microsoft\/Phi-4-reasoning\" target=\"_blank\" rel=\"noopener\">right here<\/a> and <a href=\"https:\/\/huggingface.co\/microsoft\/Phi-4-reasoning-plus\" target=\"_blank\" rel=\"noopener\">right here<\/a>.<\/p>\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/04\/image-9-1024x416.webp\" alt=\"A graph of different colored bars\" class=\"wp-image-40129\"\/><figcaption class=\"wp-element-caption\">Determine 1.\u202fPhi-4-reasoning efficiency throughout consultant reasoning benchmarks spanning mathematical and scientific reasoning. We illustrate the efficiency features from reasoning-focused post-training of Phi-4 through Phi-4-reasoning (SFT) and Phi-4-reasoning-plus (SFT+RL), alongside a consultant set of baselines from two mannequin households: open-weight fashions from DeepSeek together with DeepSeek R1 (671B Combination-of-Specialists) and its distilled dense variant DeepSeek-R1 Distill Llama 70B, and OpenAI\u2019s proprietary frontier fashions o1-mini and o3-mini. Phi-4-reasoning and Phi-4-reasoning-plus constantly outperform the bottom mannequin Phi-4 by important margins, exceed DeepSeek-R1 Distill Llama 70B (5x bigger)\u202fand show aggressive efficiency in opposition to considerably bigger fashions resembling Deepseek-R1.<\/figcaption><\/figure>\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" alt=\"A graph of numbers and a number of people\" class=\"wp-image-40018 webp-format\" srcset=\"https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/05\/image-2-1024x359.webp 1024w, https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/05\/image-2-300x105.webp 300w, https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/05\/image-2-768x269.webp 768w, https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/05\/image-2-1536x539.webp 1536w, https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/05\/image-2.webp 1600w\" src=\"https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/05\/image-2-1024x359.webp\"\/><figcaption class=\"wp-element-caption\">Determine 2. Accuracy of fashions throughout general-purpose benchmarks for: lengthy enter context\u00a0QA (FlenQA), instruction following (IFEval), Coding (HumanEvalPlus), information &amp; language understanding (MMLUPro), security detection (ToxiGen), and different normal expertise (ArenaHard and PhiBench).\u00a0<\/figcaption><\/figure>\n<p class=\"wp-block-paragraph\">Phi-4-reasoning fashions introduce a serious enchancment over Phi-4, surpass bigger fashions like DeepSeek-R1-Distill-70B and strategy Deep-Search-R1 throughout varied reasoning and normal capabilities, together with math, coding, algorithmic downside fixing, and planning. The <a href=\"https:\/\/aka.ms\/phi-reasoning\/techreport\" target=\"_blank\" rel=\"noreferrer noopener\">technical report<\/a> offers in depth quantitative proof of those enhancements by various reasoning duties.<\/p>\n<p><h2 class=\"wp-block-heading\" id=\"phi-4-mini-reasoning\"><span class=\"ez-toc-section\" id=\"Phi-4-mini-reasoning\"><\/span>Phi-4-mini-reasoning<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<\/p>\n<div class=\"wp-block-group is-layout-constrained wp-block-group-is-layout-constrained\">\n<p class=\"wp-block-paragraph\"><strong>Phi-4-mini-reasoning<\/strong> is designed to fulfill the demand for a compact reasoning mannequin. This transformer-based language mannequin is optimized for mathematical reasoning, offering high-quality, step-by-step downside fixing in environments with constrained computing or latency. Effective-tuned with artificial knowledge generated by Deepseek-R1 mannequin, Phi-4-mini-reasoning balances effectivity with superior reasoning potential. It\u2019s ultimate for academic functions, embedded tutoring, and light-weight deployment on edge or cellular techniques, and is skilled on over a million various math issues spanning a number of ranges of problem from center faculty to Ph.D. stage.\u00a0Check out the mannequin on <a href=\"https:\/\/huggingface.co\/microsoft\/Phi-4-mini-reasoning\/blob\/main\/Phi-4-Mini-Reasoning.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">Azure AI Foundry<\/a> or <a href=\"https:\/\/aka.ms\/phi4-mini-reasoning\/hf\" target=\"_blank\" rel=\"noopener\">HuggingFace<\/a> immediately.<\/p>\n<\/div>\n<figure class=\"wp-block-image aligncenter size-large\"><img decoding=\"async\" src=\"https:\/\/azure.microsoft.com\/en-us\/blog\/wp-content\/uploads\/2025\/04\/Screenshot-2025-04-30-193715-1024x328.webp\" alt=\"A graph of numbers and a number of marks\" class=\"wp-image-40226\"\/><figcaption class=\"wp-element-caption\">Determine 3. The graph compares the efficiency of assorted fashions on widespread math benchmarks for lengthy sentence era. Phi-4-mini-reasoning outperforms its base mannequin on lengthy sentence era throughout every analysis, in addition to bigger fashions like OpenThinker-7B, Llama-3.2-3B-instruct, DeepSeek-R1-Distill-Qwen-7B, DeepSeek-R1-Distill-Llama-8B, and Bespoke-Stratos-7B. Phi-4-mini-reasoning is corresponding to OpenAI o1-mini throughout math benchmarks, surpassing the mannequin\u2019s efficiency throughout Math-500 and GPQA Diamond evaluations. As seen above, Phi-4-mini-reasoning with 3.8B parameters outperforms fashions of over twice its dimension.\u202f<\/figcaption><\/figure>\n<p class=\"wp-block-paragraph\">For extra details about the mannequin, learn the\u202f<a href=\"https:\/\/arxiv.org\/pdf\/2504.21233\" target=\"_blank\" rel=\"noreferrer noopener\">technical report<\/a> that gives extra quantitative insights.<\/p>\n<p class=\"wp-block-paragraph\">Phi\u2019s evolution during the last yr has regularly pushed this envelope of high quality vs. dimension, increasing the household with new options to deal with various wants.\u00a0Throughout the dimensions of Home windows 11 units, these fashions can be found to run regionally on CPUs and GPUs.<\/p>\n<p class=\"wp-block-paragraph\">As Home windows works in direction of creating a brand new kind of PC, Phi fashions have turn out to be an integral a part of Copilot+ PCs with the NPU-optimized <a href=\"https:\/\/blogs.windows.com\/windowsexperience\/2024\/12\/06\/phi-silica-small-but-mighty-on-device-slm\/\" target=\"_blank\" rel=\"noreferrer noopener\">Phi Silica variant<\/a>. This extremely environment friendly and OS-managed model of Phi is designed to be preloaded in reminiscence, and obtainable with blazing quick time to first token responses, and energy environment friendly token throughput so it may be concurrently invoked with different functions working in your PC.<\/p>\n<p class=\"wp-block-paragraph\">It&#8217;s utilized in core experiences like <a href=\"https:\/\/support.microsoft.com\/en-us\/windows\/click-to-do-do-more-with-what-s-on-your-screen-6848b7d5-7fb0-4c43-b08a-443d6d3f5955\" target=\"_blank\" rel=\"noreferrer noopener\">Click on to Do<\/a>, offering helpful textual content intelligence instruments for any content material in your display screen, and is obtainable as <a href=\"https:\/\/learn.microsoft.com\/en-us\/windows\/ai\/apis\/phi-silica?tabs=csharp0%2Ccsharp1%2Ccsharp2%2Ccsharp3\" target=\"_blank\" rel=\"noreferrer noopener\">developer APIs<\/a> to be readily built-in into functions\u2014already being utilized in a number of productiveness functions like Outlook, providing its Copilot abstract options offline.\u00a0These small however mighty fashions have already been optimized and built-in for use throughout a number of functions throughout the breadth of our PC ecosystem.\u00a0The Phi-4-reasoning and Phi-4-mini-reasoning fashions leverage the low-bit optimizations for Phi Silica and will likely be obtainable to run quickly on Copilot+ PC NPUs.<\/p>\n<h2 class=\"wp-block-heading\" id=\"safety-and-microsoft-s-approach-to-responsible-ai\"><span class=\"ez-toc-section\" id=\"Security_and_Microsoft%E2%80%99s_strategy_to_accountable_AI\"><\/span>Security and Microsoft\u2019s strategy to accountable AI\u00a0<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p class=\"wp-block-paragraph\">At Microsoft, <a href=\"https:\/\/www.microsoft.com\/en-us\/ai\/responsible-ai?msockid=2e923f4e6e1064c017fe2d466fa365a3\" target=\"_blank\" rel=\"noreferrer noopener\">accountable AI<\/a> is a basic precept guiding the event and deployment of AI techniques, together with our Phi fashions. Phi fashions are developed in accordance with Microsoft AI ideas: accountability, transparency, equity, reliability and security, privateness and safety, and inclusiveness.\u00a0<\/p>\n<p class=\"wp-block-paragraph\">The Phi household of fashions has adopted a strong security post-training strategy, leveraging a mixture of Supervised Effective-Tuning (SFT), Direct Desire Optimization (DPO), and Reinforcement Studying from Human Suggestions (RLHF) methods. These strategies make the most of varied datasets, together with publicly obtainable datasets centered on helpfulness and harmlessness, in addition to varied safety-related questions and solutions. Whereas the Phi household of fashions is designed to carry out a variety of duties successfully, you will need to acknowledge that each one AI fashions could exhibit limitations. To higher perceive these limitations and the measures in place to deal with them, please confer with the mannequin playing cards under, which give detailed data on accountable AI practices and tips.<\/p>\n<h2 class=\"wp-block-heading\" id=\"learn-more-here\"><span class=\"ez-toc-section\" id=\"Be_taught_extra_right_here\"><\/span>Be taught extra right here:\u00a0<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<\/p><\/div>\n<p><script>\n\t\tfunction facebookTracking() {\n\t\t\t!function(f,b,e,v,n,t,s){if(f.fbq)return;n=f.fbq=function(){n.callMethod?\n\t\t\t\tn.callMethod.apply(n,arguments):n.queue.push(arguments)};if(!f._fbq)f._fbq=n;\n\t\t\t\tn.push=n;n.loaded=!0;n.version='2.0';n.queue=[];t=b.createElement(e);t.async=!0;\n\t\t\t\tt.src=v;t.type=\"ms-delay-type\";t.setAttribute('data-ms-type','text\/javascript');\n\t\t\t\ts=b.getElementsByTagName(e)[0];s.parentNode.insertBefore(t,s)}(window,\n\t\t\t\tdocument,'script','https:\/\/connect.facebook.net\/en_US\/fbevents.js');\n\t\t\tfbq('init', '1770559986549030');\n\t\t\t\t\t\tfbq('track', 'PageView');\n\t\t\t\t\t}\n\t<\/script><br \/>\n<br \/><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Microsoft continues so as to add to the dialog by unveiling its latest fashions, Phi-4-reasoning, Phi-4-reasoning-plus, and Phi-4-mini-reasoning.\u00a0 A brand new period of AI\u00a0 One yr in the past, Microsoft launched small language fashions (SLMs) to prospects with the discharge of Phi-3 on Azure AI Foundry, leveraging analysis on SLMs to increase the vary of [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":7023,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[22],"tags":[],"class_list":{"0":"post-7021","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-iot"},"_links":{"self":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/7021","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=7021"}],"version-history":[{"count":1,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/7021\/revisions"}],"predecessor-version":[{"id":7022,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/posts\/7021\/revisions\/7022"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=\/wp\/v2\/media\/7023"}],"wp:attachment":[{"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=7021"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=7021"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aireviewirush.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=7021"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}