{"id":15992,"date":"2026-11-17T11:04:25","date_gmt":"2026-11-17T02:04:25","guid":{"rendered":"https:\/\/localmodelwatch.tsuchitsuchi.com\/?p=15992"},"modified":"2026-11-17T11:13:44","modified_gmt":"2026-11-17T02:13:44","slug":"unsloth-v0-1-917-beta-released","status":"publish","type":"post","link":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/unsloth-v0-1-917-beta-released\/","title":{"rendered":"Unsloth v0.1.917-beta Released with Major Performance Boosts"},"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\/unslothai\/unsloth\">unslothai\/unsloth<\/a><\/td>\n<\/tr>\n<tr>\n<td>Version<\/td>\n<td><a href=\"https:\/\/github.com\/unslothai\/unsloth\/releases\/tag\/v0.1.917-beta\">v0.1.917-beta<\/a><\/td>\n<\/tr>\n<tr>\n<td>Published<\/td>\n<td>2026-11-12<\/td>\n<\/tr>\n<tr>\n<td>License<\/td>\n<td>Apache-2.0<\/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>Unsloth is a UI tool for running and training LLMs and diffusion models locally. In this update, Unsloth v0.1.917-beta has been released.<\/p>\n<p>The most significant change in this release is the substantial performance improvement. In particular, the operation of UI elements such as streaming, scrolling, menus, and dialogs has been enhanced, making them less prone to performance degradation even during long conversations. It also includes support for Qwen3.8-27B Dynamic v3.1 GGUF.<\/p>\n<h2>Key Changes<\/h2>\n<h3>Chat and UI Improvements<\/h3>\n<p>Chat streaming performance has been improved. Even as conversations grow longer, streaming performance is designed not to degrade. Additionally, scrolling behavior, menus, and dialog loading speeds have been enhanced. This allows users to experience smooth operation even during long dialogues.<\/p>\n<h3>Image Generation Efficiency and New Features<\/h3>\n<p>In FLUX.2 and FLUX.3, VRAM usage during image decoding has been successfully reduced by 36\u201343% while maintaining the same speed. Furthermore, an image generation queue feature has been added, allowing users to queue the next generation while another image is rendering. In addition, INT8 and FP8 image models can now be loaded on older NVIDIA GPUs, including the T4.<\/p>\n<h3>Training Speedups<\/h3>\n<p>When using FastModel models, the speed of <code>model.generate()<\/code> on T4 GPUs has improved by 2\u20138x by default. This is a massive benefit for users running and fine-tuning models locally using T4 GPUs.<\/p>\n<h3>Enhanced Coding Agents<\/h3>\n<p>For Mac users, a feature has been added to run coding agents like Claude Code and Codex using MLX models. This is available through the <code>unsloth start<\/code> command. A new Unsloth MCP preset also allows coding agents to access Unsloth&#8217;s tools.<\/p>\n<h3>Video Generation Improvements<\/h3>\n<p>In Cosmos3-Nano, a feature has been added to generate continuations from the first or last frame of a video clip. Video columns are also now properly supported when training vision models.<\/p>\n<h2>Supported Models and Hardware<\/h2>\n<p>Qwen3.8-27B Dynamic v3.1 GGUF is now supported as a new model architecture. This makes it possible to run this model locally using Unsloth.<\/p>\n<p>In the field of image generation models, the following updates have been made:<br \/>\n&#8211; VRAM usage during image decoding for FLUX.2 and FLUX.3 has been reduced by 36\u201343%.<br \/>\n&#8211; INT8 and FP8 image models can now be loaded on older NVIDIA GPUs, including the T4.<br \/>\n&#8211; LTX-2.5 FP8 now runs on 24 GB GPUs.<br \/>\n&#8211; Cosmos3-Nano can now generate continuations from the first or last frame of a video clip.<\/p>\n<p>On the training side, Ulysses Context Parallelism now supports Kimi-Linear, DeepSeek, MiniMax, and Gemma 4 text models.<\/p>\n<h2>How to Get It<\/h2>\n<p>Here are the steps to install from the latest nightly source for developers.<\/p>\n<h3>macOS, Linux, and WSL<\/h3>\n<pre><code class=\"language-bash\">git clone https:\/\/github.com\/unslothai\/unsloth\ncd unsloth\n.\/install.sh --local\nunsloth studio -p 8888\n<\/code><\/pre>\n<h3>Windows PowerShell<\/h3>\n<pre><code class=\"language-powershell\">git clone https:\/\/github.com\/unslothai\/unsloth.git\ncd unsloth\nSet-ExecutionPolicy -Scope Process -ExecutionPolicy RemoteSigned\n.\\install.ps1 --local\nunsloth studio -p 8888\n<\/code><\/pre>\n<p>To update an existing environment, run <code>git pull<\/code> within the repository and then execute the installation script again.<\/p>\n<p><!-- lmw:related --><\/p>\n<h2>Related Articles<\/h2>\n<ul>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/unsloth-v0-1-916-beta-released\/\">Unsloth v0.1.916-beta Released: Performance &amp; Multi-GPU Updates<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/unsloth-v0-1-915-beta-released\/\">Unsloth v0.1.915-beta Released with Browser Use &amp; Qwen3.8<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/unsloth-v0-1-907-beta-released\/\">Unsloth v0.1.907-beta Released with Major Speed and UI Upgrades<\/a><\/li>\n<li><a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/2026\/11\/17\/unsloth-v0-1-918-beta-release\/\">Unsloth v0.1.918-beta Adds Computer Use, MCP Game Engine 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-unsloth-en\/\">Unsloth overview and release history (89 releases tracked)<\/a><\/li>\n<li><strong>Other quantization, model formats and fine-tuning<\/strong> \u2192 <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-ggml-en\/\">ggml<\/a> \/ <a href=\"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/engine-exllamav3-en\/\">ExLlamaV3<\/a><\/li>\n<\/ul>\n<p><!-- \/lmw:next-steps --><\/p>\n<h2>Sources<\/h2>\n<ul>\n<li><a href=\"https:\/\/github.com\/unslothai\/unsloth\/releases\/tag\/v0.1.917-beta\">unslothai\/unsloth Dynamic v3.1 GGUFs + Performance Fixes<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Unsloth v0.1.917-beta is out, bringing UI improvements, VRAM reductions for FLUX, faster training on T4 GPUs, and new model support.<\/p>\n","protected":false},"author":1,"featured_media":15991,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[105],"tags":[1712,163,4916,522,509,1547],"class_list":["post-15992","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-engines-and-tools","tag-flux-en","tag-gguf-en","tag-localllm-en","tag-qwen3-8-en","tag-unsloth-en","tag-verified"],"lang":"en","translations":{"en":15992,"ja":15990},"_links":{"self":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/15992","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=15992"}],"version-history":[{"count":2,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/15992\/revisions"}],"predecessor-version":[{"id":18100,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/posts\/15992\/revisions\/18100"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media\/15991"}],"wp:attachment":[{"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/media?parent=15992"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/categories?post=15992"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/localmodelwatch.tsuchitsuchi.com\/en\/wp-json\/wp\/v2\/tags?post=15992"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}