{"id":4889,"date":"2026-09-26T22:26:51","date_gmt":"2026-09-26T13:26:51","guid":{"rendered":"https:\/\/localmodelwatch.tsuchitsuchi.com\/publisher-ifm-en\/"},"modified":"2026-09-27T06:52:24","modified_gmt":"2026-09-26T21:52:24","slug":"publisher-ifm-en","status":"publish","type":"page","link":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publisher-ifm-en\/","title":{"rendered":"IFM (MBZUAI) Open Models: Model List, Licenses and Articles"},"content":{"rendered":"<h2>About IFM (MBZUAI)<\/h2>\n<p><strong>IFM (Institute of Foundation Models)<\/strong> is an AI research lab launched in May 2025 by MBZUAI (Mohamed bin Zayed University of Artificial Intelligence) in the United Arab Emirates. According to its <a href=\"https:\/\/huggingface.co\/IFM\">IFM organization on Hugging Face<\/a>, it operates from Abu Dhabi, Silicon Valley and Paris and is dedicated to &#8220;open and independent development of frontier-class foundation models&#8221;.<\/p>\n<p>IFM stands out for <strong>&#8220;fully open&#8221; models that release not only weights but training data, recipes and training code<\/strong>. As LLM360 it previously released models under a policy of publishing all checkpoints, code and data.<\/p>\n<h2>What It Releases<\/h2>\n<ul>\n<li><strong>K2-Horizon is its latest family, with sizes from 0.9B to 375B<\/strong> (September 2026): dense models at 0.9B, 3.7B, 7B and 32B, the 36B-A4B MoVA, and a 375B-A23B MoE. According to the 32B model card, the native context is 512K tokens, and the training data, recipe and training code will be made public.<\/li>\n<li><strong>It releases intermediate checkpoints too<\/strong>, so that capability changes can be studied across training. Note that the results on the 32B model card are for stage 1 of training; stage 2 results will follow.<\/li>\n<li>Earlier families include the reasoning-focused K2-Think, the 65B K2, and the 7B Amber and Crystal. The organization card also lists Jais, an Arabic LLM, among IFM&#8217;s models.<\/li>\n<\/ul>\n<h2>Licenses<\/h2>\n<p>K2-Horizon and K2-Think are <strong>Apache-2.0<\/strong>, broadly usable including commercially. The guru 7B and 32B models, however, are non-commercial CC BY-NC 4.0, and MegaMath uses the license of its Llama 3.2 base. This summary is not legal advice; read each model&#8217;s license text before use.<\/p>\n<h2>Running Them Locally<\/h2>\n<ul>\n<li><strong>K2-Horizon has official GGUF builds.<\/strong> IFM itself publishes GGUF files for 0.9B, 3.7B, 7B, 32B and MoVA-36B-A4B. The 32B and 375B have NVFP4 builds, and many sizes have FP8 builds.<\/li>\n<li><strong>It is an easy family to pick a size from.<\/strong> On a consumer GPU the 7B or 32B GGUF, on a laptop the 0.9B or 3.7B, are the candidates. For which quantization fits which GPU, see the memory table on the family page.<\/li>\n<\/ul>\n<p><em>Sources: <a href=\"https:\/\/huggingface.co\/IFM\">the IFM organization card<\/a>; the model card of <a href=\"https:\/\/huggingface.co\/IFM\/K2-Horizon-32B\">IFM\/K2-Horizon-32B<\/a>; the model list of the IFM organization on Hugging Face (all as of 2026-09-26).<\/em><\/p>\n<h2>Our Coverage and Data<\/h2>\n<p>Local Model Watch has published 3 article(s) on IFM (MBZUAI)&#8217;s own models, 0 on third-party fine-tunes and quantized builds based on them, and 0 from its official blog. The model list below is collected by code from Hugging Face every day. Part of our <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/publishers-en\/\">publisher index<\/a>; model families link to their <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/models-en\/\">family pages<\/a>.<\/p>\n<h2>At a Glance<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Item<\/th>\n<th>Value<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Hugging Face organization<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\">IFM<\/a><\/td>\n<\/tr>\n<tr>\n<td>Models on Hugging Face<\/td>\n<td>40<\/td>\n<\/tr>\n<tr>\n<td>Articles on its own models<\/td>\n<td>3<\/td>\n<\/tr>\n<tr>\n<td>Articles on derived models<\/td>\n<td>0<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<h2>Recent Models on Hugging Face<\/h2>\n<p>The publisher lists <strong>40 models<\/strong> in total on Hugging Face. Below are up to 15 of the most recent generative models (text, image, video, audio and so on), the scope of this site, per organization. Classifiers, feature extractors, research components and repositories without a model card are omitted, and the publisher&#8217;s own quantized builds and helper variants are folded into the row of the original model. Parameter counts are left blank for repositories that pack weights into integer types, where Hugging Face&#8217;s count is not the real parameter count.<\/p>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<p class=\"lmw-table-hint\" style=\"margin:0 0 4px;font-size:0.85em;opacity:0.7;\">\u2192 Scroll horizontally to see all columns<\/p>\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Released<\/th>\n<th>Model<\/th>\n<th>Task<\/th>\n<th>Parameters<\/th>\n<th>License<\/th>\n<th>Official variants<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>2026-09-03<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/K2-Horizon-0.9B-Uno\">IFM\/K2-Horizon-0.9B-Uno<\/a><\/td>\n<td>text generation<\/td>\n<td>\u2014<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2026-09-02<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/K2-Horizon-7B-Uno\">IFM\/K2-Horizon-7B-Uno<\/a><\/td>\n<td>text generation<\/td>\n<td>\u2014<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2026-09-02<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/K2-Horizon-375B-A23B\">IFM\/K2-Horizon-375B-A23B<\/a><\/td>\n<td>text generation<\/td>\n<td>379.2B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>NVFP4, FP8<\/td>\n<\/tr>\n<tr>\n<td>2026-09-02<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/K2-Horizon-7B\">IFM\/K2-Horizon-7B<\/a><\/td>\n<td>text generation<\/td>\n<td>9.0B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>GGUF, FP8<\/td>\n<\/tr>\n<tr>\n<td>2026-09-02<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/K2-Horizon-3.7B\">IFM\/K2-Horizon-3.7B<\/a><\/td>\n<td>text generation<\/td>\n<td>5.1B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>GGUF<\/td>\n<\/tr>\n<tr>\n<td>2026-09-02<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/K2-Horizon-32B\">IFM\/K2-Horizon-32B<\/a><\/td>\n<td>text generation<\/td>\n<td>34.8B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>NVFP4, GGUF, FP8<\/td>\n<\/tr>\n<tr>\n<td>2026-09-02<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/K2-Horizon-MoVA-36B-A4B\">IFM\/K2-Horizon-MoVA-36B-A4B<\/a><\/td>\n<td>text generation<\/td>\n<td>37.4B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>FP8, GGUF<\/td>\n<\/tr>\n<tr>\n<td>2026-09-02<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/K2-Horizon-0.9B\">IFM\/K2-Horizon-0.9B<\/a><\/td>\n<td>text generation<\/td>\n<td>1.1B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>GGUF<\/td>\n<\/tr>\n<tr>\n<td>2026-01-09<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/K2-Think-V2\">IFM\/K2-Think-V2<\/a><\/td>\n<td>text generation<\/td>\n<td>72.6B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2025-09-08<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/K2-Think\">IFM\/K2-Think<\/a><\/td>\n<td>text generation<\/td>\n<td>32.8B<\/td>\n<td><code>apache-2.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2025-05-15<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/guru-32B\">IFM\/guru-32B<\/a><\/td>\n<td>text generation<\/td>\n<td>32.8B<\/td>\n<td><code>cc-by-nc-4.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2025-05-15<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/guru-7B\">IFM\/guru-7B<\/a><\/td>\n<td>text generation<\/td>\n<td>7.6B<\/td>\n<td><code>cc-by-nc-4.0<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2025-04-03<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/MegaMath-Llama-3.2-1B\">IFM\/MegaMath-Llama-3.2-1B<\/a><\/td>\n<td>text generation<\/td>\n<td>1.2B<\/td>\n<td><code>llama3.2<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2025-04-03<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/MegaMath-Llama-3.2-3B\">IFM\/MegaMath-Llama-3.2-3B<\/a><\/td>\n<td>text generation<\/td>\n<td>3.2B<\/td>\n<td><code>llama3.2<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<tr>\n<td>2024-06-27<\/td>\n<td><a href=\"https:\/\/huggingface.co\/IFM\/CrystalChat-7B-Web2Code\">IFM\/CrystalChat-7B-Web2Code<\/a><\/td>\n<td>text generation<\/td>\n<td>\u2014<\/td>\n<td><code>mit<\/code><\/td>\n<td>\u2014<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><em>From the Hugging Face API. Last changed 2026-09-27 (JST). Collected by code.<\/em><\/p>\n<h2>Licenses of Its Own Models (as recorded in our articles)<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>License<\/th>\n<th>Commercial use<\/th>\n<th>Models<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><code>apache-2.0<\/code><\/td>\n<td>Commercial use allowed<\/td>\n<td>IFM\/K2-Horizon-32B-NVFP4, IFM\/K2-Horizon-375B-A23B-NVFP4, IFM\/K2-Horizon-MoVA-36B-A4B-GGUF<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><em>The license can differ from model to model even within one publisher. This is not legal advice; check each model&#8217;s license text before use.<\/em><\/p>\n<h2>By Model Family<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<p class=\"lmw-table-hint\" style=\"margin:0 0 4px;font-size:0.85em;opacity:0.7;\">\u2192 Scroll horizontally to see all columns<\/p>\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Family<\/th>\n<th>Own-model articles<\/th>\n<th>Derived-model articles<\/th>\n<th>Smallest VRAM tier<\/th>\n<th>Latest<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/model-ifm-k2-horizon-en\/\">K2-Horizon<\/a><\/td>\n<td>3<\/td>\n<td>0<\/td>\n<td>32GB<\/td>\n<td>2026-09-23<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<h2>Articles on Its Own Models<\/h2>\n<div class=\"lmw-table-scroll\" tabindex=\"0\" style=\"overflow-x:auto;-webkit-overflow-scrolling:touch;max-width:100%;\">\n<table style=\"width:max-content;min-width:100%;border-collapse:collapse;\">\n<thead>\n<tr>\n<th>Published<\/th>\n<th>Model<\/th>\n<th>Type<\/th>\n<th>Article<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>2026-09-23<\/td>\n<td>IFM\/K2-Horizon-32B-NVFP4<\/td>\n<td>New Models<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/23\/ifm-k2-horizon-32b-nvfp4-released\/\">K2-Horizon-32B-NVFP4 Long-Context Reasoning Model: 32GB+ VRAM<\/a><\/td>\n<\/tr>\n<tr>\n<td>2026-09-23<\/td>\n<td>IFM\/K2-Horizon-375B-A23B-NVFP4<\/td>\n<td>New Models<\/td>\n<td><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/23\/ifm-k2-horizon-375b-a23b-nvfp4-released-2\/\">K2-Horizon-375B-A23B-NVFP4 Text Generation Model: ~257GB Memory<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><em>Last updated 2026-09-27 (JST). The explanation at the top of this page was written with the help of AI from the primary sources it cites. The tables and lists under &#8220;Our Coverage and Data&#8221; are assembled by code from our article log and the Hugging Face API.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"<p>About IFM (MBZUAI) IFM (Institute of Foundation Models) is an AI research lab launched in May 2025 by MBZUAI (Mohamed bin Zayed University of Artificial Intelligence) in the United Arab Emirates. According to its IFM organization on Hugging Face, it operates from Abu Dhabi, Silicon Valley and Paris and is dedicated to &#8220;open and independent [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-4889","page","type-page","status-publish","hentry"],"lang":"en","translations":{"en":4889,"ja":4888},"_links":{"self":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/pages\/4889","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/comments?post=4889"}],"version-history":[{"count":1,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/pages\/4889\/revisions"}],"predecessor-version":[{"id":5038,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/pages\/4889\/revisions\/5038"}],"wp:attachment":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media?parent=4889"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}