{"id":21167,"date":"2026-11-21T02:13:11","date_gmt":"2026-11-20T17:13:11","guid":{"rendered":"https:\/\/localmodelwatch.tsuchitsuchi.com\/2026\/11\/21\/localai-v5-0-1-release-2\/"},"modified":"2026-11-21T02:13:11","modified_gmt":"2026-11-20T17:13:11","slug":"localai-v5-0-1-release-2","status":"publish","type":"post","link":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/21\/localai-v5-0-1-release-2\/","title":{"rendered":"LocalAI v5.0.1 Released: Native Windows Support and ROCm Enhancements"},"content":{"rendered":"<p><!-- lmw:facts --><\/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>Repository<\/td>\n<td><a href=\"https:\/\/github.com\/mudler\/LocalAI\">mudler\/LocalAI<\/a><\/td>\n<\/tr>\n<tr>\n<td>Version<\/td>\n<td><a href=\"https:\/\/github.com\/mudler\/LocalAI\/releases\/tag\/v5.0.1\">v5.0.1<\/a><\/td>\n<\/tr>\n<tr>\n<td>Published<\/td>\n<td>2026-11-21<\/td>\n<\/tr>\n<tr>\n<td>License<\/td>\n<td>MIT<\/td>\n<\/tr>\n<tr>\n<td>Source type<\/td>\n<td>Primary source (the publisher itself)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><em>Values determined by this site&#8217;s code when the information was collected. Dates are JST.<\/em><\/p>\n<p><!-- \/lmw:facts --><\/p>\n<h2>Overview<\/h2>\n<p>LocalAI, an open-source AI engine, has released version <code>v5.0.1<\/code>. LocalAI is MIT-licensed software developed in Go that allows users to run generative models for LLM, vision, audio, image, and video tasks on a variety of hardware, including environments without a GPU.<\/p>\n<p>The most significant change for users in this release is the addition of native support for Windows environments. With the server itself and the <code>llama-cpp<\/code> backend now natively supporting <code>windows\/amd64<\/code>, deployment and operation on Windows have become significantly easier.<\/p>\n<h2>Key Changes<\/h2>\n<h3>Native Windows Support Added<\/h3>\n<p>Native support has been added for Windows environments, which previously often required additional configurations or container environments, through the server itself and the <code>llama-cpp<\/code> backend for <code>windows\/amd64<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/11429\">PR #11429<\/a>). This greatly improves convenience for users who want to run local LLMs and inference engines directly on Windows. No special configuration migration is required for the update, and direct execution on Windows is now possible.<\/p>\n<h3>Improved Model Downloader Safety<\/h3>\n<p>The model download process has been modified to retain existing old model files until the new model file is correctly verified (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13308\">PR #13308<\/a>). This prevents the risk of previous model files being unexpectedly corrupted or lost in the event of a network disconnection or error during the download. This is a reassuring change for users who frequently update or add models.<\/p>\n<h3>Enhanced ROCm (AMD GPU) Support and Bug Fixes<\/h3>\n<p>Several enhancements and fixes have been implemented for environments using AMD GPUs. The hipblas image is now built based on ROCm 10.1 by default (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13169\">PR #13169<\/a>), and <code>*_TENSILE_LIBPATH<\/code> has been corrected to properly point to the folder of the installed GPU (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13321\">PR #13321<\/a>). Additionally, the privacy-filter now builds a ROCm (hipblas) image (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13316\">PR #13316<\/a>). Users running LocalAI in AMD GPU environments can expect resolution of runtime library path-related errors and improved compatibility.<\/p>\n<h3>Expanded Platform and Launcher Support<\/h3>\n<p>The <code>darwin\/arm64<\/code> (Apple Silicon Mac) package has been enabled in the sandbox environment (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13306\">PR #13306<\/a>). Furthermore, fixes have been applied so that the x64 version of the AppImage can start correctly even in environments without glibc (such as musl-based Linux environments) (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13340\">PR #13340<\/a>). This expands the range of operating environments for users who previously faced restrictions on certain Linux distributions or macOS.<\/p>\n<h3>Modernization of Major Inference Backends<\/h3>\n<p>Various inference libraries used internally by LocalAI have been updated. Specifically, <code>ggml-org\/llama.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13284\">PR #13284<\/a>), <code>ggml-org\/whisper.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13283\">PR #13283<\/a>), <code>ikawrakow\/ik_llama.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13290\">PR #13290<\/a>), <code>mudler\/timestral.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13286\">PR #13286<\/a>), <code>0xShug0\/audio.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13282\">PR #13282<\/a>), <code>CrispStrobe\/CrispASR<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13288\">PR #13288<\/a>), and <code>ServeurpersoCom\/omnivoice.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13287\">PR #13287<\/a>) have all been updated to their latest commits. This affects all users running text generation, speech recognition, or text-to-speech models, applying the latest fixes from each upstream project.<\/p>\n<h3>UI and Desktop Feature Improvements<\/h3>\n<p>Enhancements have been made to the user interface and desktop application features. A desktop notification function for the server inbox has been implemented (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13298\">PR #13298<\/a>), and &#8220;Custom&#8221; labeling and domain display have been added to network options in the UI (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13294\">PR #13294<\/a>). Additionally, Bengali (bn) translation has been added for the React UI and add-ons (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13304\">PR #13304<\/a>). This improves usability for users operating primarily via the desktop app or Web UI.<\/p>\n<h2>Supported Models and Hardware<\/h2>\n<p>LocalAI v5.0.1 expands the range of supported hardware environments and backends.<\/p>\n<p>First, native support for Windows has been implemented. With the addition of the <code>llama-cpp<\/code> backend for the <code>windows\/amd64<\/code> architecture, <code>llama.cpp<\/code>-based model inference can now be executed directly on Windows without using containers or similar wrappers (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/11429\">PR #11429<\/a>).<\/p>\n<p>Support for AMD GPU (ROCm) environments has also been updated. The hipblas image is now built based on ROCm 10.1 by default (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13169\">PR #13169<\/a>), and a ROCm (hipblas) image has been added to the build targets for the privacy-filter component (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13316\">PR #13316<\/a>). Furthermore, <code>*_TENSILE_LIBPATH<\/code> has been corrected to properly point to the installed GPU folder (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13321\">PR #13321<\/a>). This enables more stable model execution on servers and PCs equipped with AMD GPUs.<\/p>\n<p>Additionally, support for Apple Silicon (macOS) has been improved by enabling the <code>darwin\/arm64<\/code> package in the sandbox environment (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13306\">PR #13306<\/a>).<\/p>\n<p>Regarding the internal inference backends, several major libraries have been updated to their latest versions. Specifically, the following repositories have been updated to their latest commits:<\/p>\n<ul>\n<li><code>ggml-org\/llama.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13284\">PR #13284<\/a>)<\/li>\n<li><code>ggml-org\/whisper.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13283\">PR #13283<\/a>)<\/li>\n<li><code>ikawrakow\/ik_llama.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13290\">PR #13290<\/a>)<\/li>\n<li><code>mudler\/timestral.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13286\">PR #13286<\/a>)<\/li>\n<li><code>0xShug0\/audio.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13282\">PR #13282<\/a>)<\/li>\n<li><code>CrispStrobe\/CrispASR<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13288\">PR #13288<\/a>)<\/li>\n<li><code>ServeurpersoCom\/omnivoice.cpp<\/code> (<a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13287\">PR #13287<\/a>)<\/li>\n<\/ul>\n<p>These backend updates allow users to utilize the latest features and fixes in various inference engines for text generation, speech recognition (Whisper, etc.), and text-to-speech models.<\/p>\n<h2>How to Get It<\/h2>\n<p>Multiple methods are provided for installing and updating LocalAI, depending on your environment. Typical procedures following the official installation guide are as follows:<\/p>\n<h3>Using the Installation Script (Linux and Apple Silicon macOS)<\/h3>\n<p>This automatically downloads the release binary for your machine, verifies the checksum, and performs the installation. Run the following command in your terminal:<\/p>\n<pre><code class=\"language-bash\">curl -sSL https:\/\/localai.io\/install.sh | sh\nlocal-ai\n<\/code><\/pre>\n<p>The installation script only handles program placement, so run the <code>local-ai<\/code> command after installation to start the server. Once started, open <code>http:\/\/localhost:8080<\/code> in your browser.<\/p>\n<p>If you want to start by specifying a model directly, run it with the model name as follows (e.g., the small model for CPU <code>gemma-4-e2b-it-qat-q4_0<\/code>):<\/p>\n<pre><code class=\"language-bash\">local-ai run gemma-4-e2b-it-qat-q4_0\n<\/code><\/pre>\n<h3>Using Containers (Docker \/ podman, etc.)<\/h3>\n<p>To run the latest image in a container environment like Docker, use the following command:<\/p>\n<pre><code class=\"language-bash\">docker run -ti --name local-ai -p 8080:8080 localai\/localai:latest\n<\/code><\/pre>\n<p>To restart an existing container that has already been created, run the following command:<\/p>\n<pre><code class=\"language-bash\">docker start -i local-ai\n<\/code><\/pre>\n<h3>Using the Desktop App<\/h3>\n<p>A desktop application is provided for macOS and Linux. It automatically handles server installation, updates, starting, and stopping, and allows you to view the web interface in a dedicated window. Download it from the <a href=\"https:\/\/github.com\/mudler\/LocalAI\/releases\/latest\">latest release page<\/a>. For detailed usage instructions, refer to the <a href=\"https:\/\/localai.io\/docs\/start\/install\/desktop\/\">Desktop App Guide<\/a>.<\/p>\n<p><!-- lmw:related --><\/p>\n<h2>Related Articles<\/h2>\n<ul>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/19\/localai-v5-0-0-release-2\/\">LocalAI v5.0.0 Released: Major Overhaul of Chat and Agent Systems<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/localai-v4-12-0-released-2\/\">LocalAI v4.12.0 Released: Web UI Refresh and Streaming Audio<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/10\/03\/localai-v4-11-0-released\/\">LocalAI v4.11.0 Released with Failover and NeMo Audio Models<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/09\/18\/localai-v4-10-0-released\/\">LocalAI v4.10.0 Released: Fleet Dashboard &amp; M5 Support<\/a><\/li>\n<\/ul>\n<p><!-- \/lmw:related --><\/p>\n<p><!-- lmw:next-steps --><\/p>\n<h2>What to Read Next<\/h2>\n<ul>\n<li><strong>Follow this tool<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-localai-en\/\">LocalAI overview and release history (140 releases tracked)<\/a><\/li>\n<li><strong>Other inference engines and runtimes<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-llama-cpp-en\/\">llama.cpp<\/a> \/ <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-ollama-en\/\">Ollama<\/a> \/ <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-vllm-en\/\">vLLM<\/a><\/li>\n<li><strong>Engines mentioned in this article<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-whisper-cpp-en\/\">whisper.cpp<\/a><\/li>\n<\/ul>\n<p><!-- \/lmw:next-steps --><\/p>\n<h2>Sources<\/h2>\n<ul>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/releases\/tag\/v5.0.1\">mudler\/LocalAI v5.0.1 Release Notes<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/11429\">PR #11429: add native Windows support (server + windows\/amd64 llama-cpp backend)<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13169\">PR #13169: build the hipblas images on ROCm 10.1 by default<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13282\">PR #13282: Update 0xShug0\/audio.cpp<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13283\">PR #13283: Update ggml-org\/whisper.cpp<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13284\">PR #13284: Update ggml-org\/llama.cpp<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13286\">PR #13286: Update mudler\/timestral.cpp<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13287\">PR #13287: Update ServeurpersoCom\/omnivoice.cpp<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13288\">PR #13288: Update CrispStrobe\/CrispASR<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13290\">PR #13290: Update ikawrakow\/ik_llama.cpp<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13306\">PR #13306: enable the darwin\/arm64 package<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13316\">PR #13316: build a ROCm (hipblas) image for privacy-filter<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/mudler\/LocalAI\/pull\/13321\">PR #13321: point *_TENSILE_LIBPATH at the folder of the installed GPU<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>LocalAI v5.0.1 is now available, featuring native Windows support, improved ROCm\/AMD GPU compatibility, and updated inference backends.<\/p>\n","protected":false},"author":1,"featured_media":21166,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[105],"tags":[163,1073,1507,1573,1547,1015],"class_list":["post-21167","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-engines-and-tools","tag-gguf-en","tag-llama-cpp-en","tag-localai-en","tag-rocm-en","tag-verified","tag-whisper-cpp-en"],"lang":"en","translations":{"en":21167,"ja":21165},"_links":{"self":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/21167","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/types\/post"}],"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=21167"}],"version-history":[{"count":0,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/21167\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media\/21166"}],"wp:attachment":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media?parent=21167"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/categories?post=21167"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/tags?post=21167"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}