# unsloth changelog > Local UI to run and train LLMs and diffusion models, including Qwen3.8, Kimi K3, MiniMax-H3, Gemma 4, DeepSeek-V4, FLUX and more. - Vendor: unslothai - Category: AI - Official site: https://unsloth.ai/docs - Tracked by: What's New (https://whatsnew.fyi/product/unsloth) - Harvested from: GitHub (unslothai/unsloth) - Entries below: 10 (newest first) What's New is an index, not a publisher: every entry below links to the vendor's own release notes, which are the authoritative source. Entries are labelled where they are hand-curated sample data, pre-releases, or drawn from a secondary source such as a developer blog. Reuse: the summaries, labels and curation here are © What's New. Quote freely with attribution and a link back; wholesale republication of the corpus is not permitted — terms: https://whatsnew.fyi/terms. The vendors' own release notes remain their publishers'. ## Releases ### v0.1.804-beta — Qwen3.8-Flash-Next + GLM-5.3-Flash - Date: 2026-08-27 - Version: v0.1.804-beta - Original notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.804-beta - Permalink: https://whatsnew.fyi/product/unsloth/releases/v0.1.804-beta - Labels: Pre-release - **added** — Support for running Qwen3.8-Flash-Next locally on 75GB RAM - **added** — Support for running GLM-5.3-Flash locally on 102GB total RAM and VRAM - **added** — Local chats resume after a disconnect instead of losing the reply - **added** — Export chats as JSONL for backups or use in other tools - **added** — Search and download embedding models directly from Hugging Face - **changed** — Inference speed improved 5x for RAM offloading - **changed** — Large GGUFs automatically split across GPU and system RAM - **changed** — Smarter GPU and RAM offloading to run larger models with less setup - **changed** — Vision chats now handle multiple images properly - **changed** — Text-to-speech models only load when actually used - **changed** — Model settings stay saved when switching chats - **changed** — Collapse tool activity by default for cleaner agent chats - **changed** — Deep Research keeps going when a provider asks it to slow down - **changed** — Model reasoning defaults now followed in CLI start command - **fixed** — Infinite repeated compaction now works - **fixed** — Linux voice recording fixed - **fixed** — NVIDIA and Wayland interface freezes fixed - **fixed** — AMD model loading crashes fixed - **fixed** — llama.cpp models now load from Windows profiles with non-English characters - **fixed** — Non-English web links now work properly as chat sources Qwen3.8-Flash-Next and GLM-5.3-Flash can now run locally in Unsloth! * Run Qwen3.8-Flash on 75GB RAM, GLM-5.3-Flash on 102GB RAM+VRAM * 5x Faster inference for RAM offloading * "Infinite" repeated compaction now works * 100+ chat, reliability and performance improvements Qwen Guide: https://unsloth.ai/docs/models/qwen3.8-next Qwen GGUFs: https://huggingface.co/unsloth/Qwen3.8-Flash-Next-GGUF GLM Guide: https://unsloth.ai/docs/models/glm-5.3-flash GLM GGUFs: https://huggingface.co/unsloth/GLM-5.3-Flash-GGUF ##### Highlights - **Qwen3.8-Flash-Next on 75GB RAM** - **GLM-5.3-Flash on 102GB total memory** - **Smarter GPU + RAM offloading** - run larger models with less setup - **Chats recover after disconnects** instead of losing the reply - **See what fits before loading** with clearer memory estimates ##### Qwen3.8-Flash-Next Qwen3.8-Flash-Next is a new 125B multimodal reasoning model and an early preview of Qwen4's architecture. - The 1-bit Unsloth Dynamic GGUF runs on 75GB RAM or unified memory. - It's 79% smaller than BF16 while retaining 80% top-1 accuracy. - Chat with text and images using up to 262K context. - Switch between None, Low, Medium and Extra High reasoning. - Preserved Thinking keeps reasoning consistent across longer chats. ##### GLM-5.3-Flash GLM-5.3-Flash is Z.ai's new 320B multimodal model, with only 18B parameters active at a time. - Run the 1-bit model on 102GB of combined RAM + VRAM. - Chat with text, images and long documents using up to 1M context. - Switch between Low, High and Max reasoning. - Stronger coding, agent and vision performance than GLM-5.2. - Recommended settings are applied automatically in Unsloth. ##### Chat + tools - Local chats resume after a disconnect instead of losing the reply. - Deep Research keeps going when a provider asks it to slow down. - Vision chats now handle multiple images properly. - Images returned by MCP tools appear directly in chat. - Export chats as JSONL for backups or use in other tools. - Adjust Auto Compaction for longer chats, or turn it off. - Collapse tool activity by default for cleaner agent chats. ##### Models + performance - Large GGUFs automatically split across GPU and system RAM. - See estimated memory usage before loading a model. - View VRAM usage directly from your downloaded models. - Model settings stay saved when switching chats. - Search and download embedding models directly from Hugging Face. - Text-to-speech models only load when you actually use them. ##### Desktop + reliability - Linux voice recording fixed. - NVIDIA + Wayland interface freezes fixed. - AMD model loading crashes fixed. - llama.cpp models now load from Windows profiles with non-English characters. - Non-English web links now work properly as chat sources. - Desktop download links always point to the latest stable release. ##### What's Changed * Bump install.sh / install.ps1 pin to unsloth>=2026.8.21 by @danielhanchen in https://github.com/unslothai/unsloth/pull/9699 * fix(studio): resolve PowerShell by absolute path in the update gate by @yzxcj797 in https://github.com/unslothai/unsloth/pull/9452 * Prevent shared preview loads from evicting the active Studio model by @NilayYadav in https://github.com/unslothai/unsloth/pull/7104 * studio: harden lockfile audit followups for #5604 by @danielhanchen in https://github.com/unslothai/unsloth/pull/5695 * Fix Linux voice recording by capturing raw PCM where WebKitGTK's MediaRecorder produces no audio by @Fizza-Mukhtar in https://github.com/unslothai/unsloth/pull/9564 * Studio: take remend 1.3.1, which stops repairing markdown that is already complete by @danielhanchen in https://github.com/unslothai/unsloth/pull/9667 * CLI: follow model reasoning default in unsloth start by @shimmyshimmer in https://github.com/unslothai/unsloth/pull/9733 * Name the encoding when the poll probe writes its artifact by @danielhanchen in https://github.co _[Truncated at 4000 characters — full notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.804-beta]_ ### v0.1.803-beta — Bug Fixes + Auto compaction + LAN Remote Access - Date: 2026-08-25 - Version: v0.1.803-beta - Original notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.803-beta - Permalink: https://whatsnew.fyi/product/unsloth/releases/v0.1.803-beta - Labels: Pre-release - **fixed** — Fixed MLX and Mac runtimes that did not run correctly - **added** — Added LAN API keyless and password-less access with keyboard shortcuts - **added** — Added XET and HTTP download toggle for clearer download progress - **fixed** — Fixed AMD bugs and 170 bug, reliability, and performance issues - **added** — Added Auto Compaction (Experimental) for longer chats beyond context limits - **added** — Added Remote and LAN Access (Preview) for easy network access without Cloudflare links - **changed** — Improved chat streaming performance, reduced UI lag, and made long conversations smoother - **added** — Added support for custom llama.cpp builds with toggles for Cache RAM, Mmap, Mlock, Checkpoints, Speculative Decoding KV Cache, and Vision On/Off - **added** — Released Unsloth Dynamic v3.0 with Qwen3.8-27B GGUFs delivering over 10% higher top-1 accuracy - **added** — Added prompt queueing and customizable keyboard shortcuts - **added** — Added edit_file tool for partial file modifications - **changed** — Improved tool calling and MCP behavior - **added** — Added Intel XPU support - **changed** — Responses API now supports structured output formatting - **changed** — Improved recovery when llama-server crashes - **added** — Added OpenCode V2 support for unsloth start - **changed** — Improved GGUF downloads to be faster - **changed** — Improved Desktop packaging and permissions - **changed** — Improved training scheduler and preprocessing - **changed** — Improved compatibility across ROCm, xFormers, and flash-attention setups Thanks for the support for Qwen3.8-27B and Unsloth Desktop! This is a bug fix release with 170+ PRs. * **MLX fixed** - Some MLX and Mac runtimes did not run correctly * **LAN API keyless / password-less** + Keyboard shortcuts * **XET / HTTP download toggle** - clearer download progress * **AMD bug fixes** + 170 bug, reliability & performance fixes ##### Features * **Auto Compaction (Experimental)** for longer chats beyond context limits * **Remote & LAN Access (Preview)** for easy network access without Cloudflare links * **Faster Chat** - Improved streaming performance, reduced UI lag, and smoother long conversations. * Support for **custom llama.cpp builds**. Toggles for Cache RAM, Mmap, Mlock, Checkpoints, Spe Decoding KV Cache, Vision On / Off * [Unsloth Dynamic v3.0](https://unsloth.ai/docs/basics/dynamic-3.0-ggufs) is released. New Qwen3.8-27B Dynamic v3.0 GGUFs deliver >10% higher top-1 accuracy compared to everyone else. Works with Unsloth. ##### Auto compaction (Experimental) You can now do long chats and surpass a model's context limit, with evicted turns remaining searchable. * Long local chats can continue past the context limit by rolling older turns out of active context. * Auto compaction creates fresh context epochs instead of permanently trimming replies. * Evicted conversations remain searchable, with improved fact recall. * Context window information is shown before chat begins. **How it works** * Entire oldest turns are removed only when needed, never mid-message. The saved transcript stays unchanged. * Evicted turns are indexed into a per-thread searchable archive using Unsloth's existing RAG pipeline (store, chunker, embedder, retrieval). Lexical search is prioritized because chat recall is usually exact matches like names, numbers, or IDs. * A recall is forced during eviction instead of relying on the model to search itself. Later retrieval uses `search_conversation`. * The archive persists across epochs, allowing future compactions to recover previously evicted context. * No summarization: it showed little benefit and added ~190s per compaction. ##### Remote & LAN access (Preview) Access Unsloth from another device on your network, managed from settings. * New dedicated Settings section for remote access. * Enable or disable LAN access without restarting. * Supports connection addresses, QR codes, and optional auto-start. * LAN access is disabled by default and requires changing the generated admin password. ##### Chat + Functionality improvements * Long chats are faster with optimized streaming and thread handling. * Projects organize chats, files, and workspaces together. * Chats remember composer settings, system prompts, and model sampling options. * Added prompt queueing and customizable keyboard shortcuts. * Added `edit_file` tool for partial file modifications. * Improved tool calling and MCP behavior. ##### Hardware + inference improvements * Support for custom llama.cpp builds. * Toggles for advanced settings - Cache RAM, Mmap, Mlock, Checkpoints, Speculative Decoding KV Cache, Vision On / Off * Better validation for missing GPU support and oversized GGUF models. * Improved MTP performance and VRAM handling. * Better compatibility across ROCm, xFormers, and flash-attention setups. * Added Intel XPU support. ##### API + serving * Responses API supports structured output formatting. * Improved recovery when llama-server crashes. * Added OpenCode V2 support for `unsloth start`. ##### Hub + Desktop + training * Hub supports more image/video workflows. * Faster and improved GGUF downloads. * Desktop packaging and permissions improved. * Training scheduler and preprocessing improvements. ##### What's Changed * Bump install.sh / install.ps1 pin to unsloth>=2026.8.19 by @danielhanchen in https://github.com/unslothai/unsloth/pull/9422 * Say what killed a UI wrapper script instead of reporting a bare 143 by @danielhan _[Truncated at 4000 characters — full notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.803-beta]_ ### v0.1.802-beta — Bug Fixes + Auto compaction + LAN Remote Access - Date: 2026-08-25 - Version: v0.1.802-beta - Original notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.802-beta - Permalink: https://whatsnew.fyi/product/unsloth/releases/v0.1.802-beta - Labels: Pre-release - **fixed** — MLX and Mac runtimes now run correctly - **added** — LAN API keyless and password-less access is now supported - **added** — XET and HTTP download toggle for clearer download progress - **fixed** — AMD bug fixes for Strix Halo and all RDNA GPUs - **added** — Auto Compaction (Experimental) for longer chats beyond context limits - **added** — Remote and LAN Access (Preview) for easy network access without Cloudflare links - **changed** — Faster Chat with improved streaming performance, reduced UI lag, and smoother long conversations - **added** — Support for custom llama.cpp builds with toggles for Cache RAM, Mmap, Mlock, Checkpoints, Speculative Decoding KV Cache, and Vision On/Off - **added** — Unsloth Dynamic v3.0 released with new Qwen3.8-27B Dynamic v3.0 GGUFs delivering >10% higher top-1 accuracy - **added** — Chats remember composer settings, system prompts, and model sampling options - **added** — Prompt queueing and customizable keyboard shortcuts - **added** — Added edit_file tool for partial file modifications - **changed** — Improved tool calling and MCP behavior - **added** — Responses API supports structured output formatting - **changed** — Improved recovery when llama-server crashes - **added** — Added OpenCode V2 support for unsloth start - **added** — Intel XPU support added - **changed** — Improved MTP performance and VRAM handling - **changed** — Better compatibility across ROCm, xFormers, and flash-attention setups Thanks for the support for Qwen3.8-27B and Unsloth Desktop! This is a bug fix release with 170+ PRs. * **MLX fixed** - Some MLX and Mac runtimes did not run correctly * **LAN API keyless / password-less** is now supported * **XET / HTTP download toggle** - clearer download progress * **AMD bug fixes** for Strix Halo, all RDNA GPUs + 170 bug fixes ##### Features * **Auto Compaction (Experimental)** for longer chats beyond context limits * **Remote & LAN Access (Preview)** for easy network access without Cloudflare links * **Faster Chat** - Improved streaming performance, reduced UI lag, and smoother long conversations. * Support for **custom llama.cpp builds**. Toggles for Cache RAM, Mmap, Mlock, Checkpoints, Spe Decoding KV Cache, Vision On / Off * [Unsloth Dynamic v3.0](https://unsloth.ai/docs/basics/dynamic-3.0-ggufs) is released. New Qwen3.8-27B Dynamic v3.0 GGUFs deliver >10% higher top-1 accuracy compared to everyone else. Works with Unsloth. ##### Auto compaction (Experimental) You can now do long chats and surpass a model's context limit, with evicted turns remaining searchable. * Long local chats can continue past the context limit by rolling older turns out of active context. * Auto compaction creates fresh context epochs instead of permanently trimming replies. * Evicted conversations remain searchable, with improved fact recall. * Context window information is shown before chat begins. **How it works** * Entire oldest turns are removed only when needed, never mid-message. The saved transcript stays unchanged. * Evicted turns are indexed into a per-thread searchable archive using Unsloth's existing RAG pipeline (store, chunker, embedder, retrieval). Lexical search is prioritized because chat recall is usually exact matches like names, numbers, or IDs. * A recall is forced during eviction instead of relying on the model to search itself. Later retrieval uses `search_conversation`. * The archive persists across epochs, allowing future compactions to recover previously evicted context. * No summarization: it showed little benefit and added ~190s per compaction. ##### Remote & LAN access (Preview) Access Unsloth from another device on your network, managed from settings. * New dedicated Settings section for remote access. * Enable or disable LAN access without restarting. * Supports connection addresses, QR codes, and optional auto-start. * LAN access is disabled by default and requires changing the generated admin password. ##### Chat + Functionality improvements * Long chats are faster with optimized streaming and thread handling. * Projects organize chats, files, and workspaces together. * Chats remember composer settings, system prompts, and model sampling options. * Added prompt queueing and customizable keyboard shortcuts. * Added `edit_file` tool for partial file modifications. * Improved tool calling and MCP behavior. ##### Hardware + inference improvements * Support for custom llama.cpp builds. * Toggles for advanced settings - Cache RAM, Mmap, Mlock, Checkpoints, Speculative Decoding KV Cache, Vision On / Off * Better validation for missing GPU support and oversized GGUF models. * Improved MTP performance and VRAM handling. * Better compatibility across ROCm, xFormers, and flash-attention setups. * Added Intel XPU support. ##### API + serving * Responses API supports structured output formatting. * Improved recovery when llama-server crashes. * Added OpenCode V2 support for `unsloth start`. ##### Hub + Desktop + training * Hub supports more image/video workflows. * Faster and improved GGUF downloads. * Desktop packaging and permissions improved. * Training scheduler and preprocessing improvements. ##### What's Changed * Bump install.sh / install.ps1 pin to unsloth>=2026.8.19 by @danielhanchen in https://github.com/unslothai/unsloth/pull/9422 * Say what killed a UI wrapper script instead of reporting a bare 143 by @danielhanc _[Truncated at 4000 characters — full notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.802-beta]_ ### v0.1.801-beta — Auto compaction (preview) + LAN Remote Access - Date: 2026-08-20 - Version: v0.1.801-beta - Original notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.801-beta - Permalink: https://whatsnew.fyi/product/unsloth/releases/v0.1.801-beta - Labels: Pre-release - **added** — Auto Compaction (Experimental) for longer chats beyond context limits with evicted turns remaining searchable - **added** — Remote & LAN Access (Preview) for easy network access without Cloudflare links - **added** — Support for custom llama.cpp builds with toggles for Cache RAM, Mmap, Mlock, Checkpoints, Speculative Decoding KV Cache, and Vision On/Off - **added** — Unsloth Dynamic v3.0 with new Qwen3.8-27B Dynamic v3.0 GGUFs delivering >10% higher top-1 accuracy - **added** — Projects feature to organize chats, files, and workspaces together - **added** — Prompt queueing and customizable keyboard shortcuts - **added** — edit_file tool for partial file modifications - **added** — Intel XPU support - **added** — Responses API supports structured output formatting - **added** — OpenCode V2 support for unsloth start - **changed** — Improved streaming performance, reduced UI lag, and smoother long conversations - **changed** — Chats now remember composer settings, system prompts, and model sampling options - **changed** — Improved tool calling and MCP behavior - **changed** — Better validation for missing GPU support and oversized GGUF models - **changed** — Improved MTP performance and VRAM handling - **changed** — Better compatibility across ROCm, xFormers, and flash-attention setups - **changed** — Faster and improved GGUF downloads - **changed** — Hub supports more image/video workflows - **changed** — Desktop packaging and permissions improved - **changed** — Training scheduler and preprocessing improvements - **fixed** — Improved recovery when llama-server crashes Thanks for the support for Qwen3.8-27B and Unsloth Desktop last week! For this release, we merged 200+ PRs to introduce many new features, fixes including: * **Auto Compaction (Experimental)** for longer chats beyond context limits * **Remote & LAN Access (Preview)** for easy network access without Cloudflare links * **Faster Chat** - Improved streaming performance, reduced UI lag, and smoother long conversations. * Support for **custom llama.cpp builds**. Toggles for Cache RAM, Mmap, Mlock, Checkpoints, Spe Decoding KV Cache, Vision On / Off * [Unsloth Dynamic v3.0](https://unsloth.ai/docs/basics/dynamic-3.0-ggufs) is released. New Qwen3.8-27B Dynamic v3.0 GGUFs deliver >10% higher top-1 accuracy compared to everyone else. Works with Unsloth. ##### Auto compaction (Experimental) You can now do long chats and surpass a model's context limit, with evicted turns remaining searchable. * Long local chats can continue past the context limit by rolling older turns out of active context. * Auto compaction creates fresh context epochs instead of permanently trimming replies. * Evicted conversations remain searchable, with improved fact recall. * Context window information is shown before chat begins. **How it works** * Entire oldest turns are removed only when needed, never mid-message. The saved transcript stays unchanged. * Evicted turns are indexed into a per-thread searchable archive using Unsloth's existing RAG pipeline (store, chunker, embedder, retrieval). Lexical search is prioritized because chat recall is usually exact matches like names, numbers, or IDs. * A recall is forced during eviction instead of relying on the model to search itself. Later retrieval uses `search_conversation`. * The archive persists across epochs, allowing future compactions to recover previously evicted context. * No summarization: it showed little benefit and added ~190s per compaction. ##### Remote & LAN access (Preview) Access Unsloth from another device on your network, managed from settings. * New dedicated Settings section for remote access. * Enable or disable LAN access without restarting. * Supports connection addresses, QR codes, and optional auto-start. * LAN access is disabled by default and requires changing the generated admin password. ##### Chat + Functionality improvements * Long chats are faster with optimized streaming and thread handling. * Projects organize chats, files, and workspaces together. * Chats remember composer settings, system prompts, and model sampling options. * Added prompt queueing and customizable keyboard shortcuts. * Added `edit_file` tool for partial file modifications. * Improved tool calling and MCP behavior. ##### Hardware + inference improvements * Support for custom llama.cpp builds. * Toggles for advanced settings - Cache RAM, Mmap, Mlock, Checkpoints, Speculative Decoding KV Cache, Vision On / Off * Better validation for missing GPU support and oversized GGUF models. * Improved MTP performance and VRAM handling. * Better compatibility across ROCm, xFormers, and flash-attention setups. * Added Intel XPU support. ##### API + serving * Responses API supports structured output formatting. * Improved recovery when llama-server crashes. * Added OpenCode V2 support for `unsloth start`. ##### Hub + Desktop + training * Hub supports more image/video workflows. * Faster and improved GGUF downloads. * Desktop packaging and permissions improved. * Training scheduler and preprocessing improvements.
| Platform | Link |
| Windows | Download |
| macOS | Download
_[Truncated at 4000 characters — full notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.801-beta]_
### v0.1.800-beta — Qwen3.8-27B
- Date: 2026-08-14
- Version: v0.1.800-beta
- Original notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.800-beta
- Permalink: https://whatsnew.fyi/product/unsloth/releases/v0.1.800-beta
- Labels: Pre-release
- **added** — Qwen3.8-27B and Qwen3.8-2.4T can now be run locally in Unsloth
- **added** — Run Qwen3.8-27B on 17GB RAM via Unsloth Dynamic GGUFs
- **added** — Fine-tune Qwen3.8-27B in Unsloth
- **added** — NVFP4 quants for Qwen3.8-27B
- **added** — Extra llama-server arguments can be passed through model settings in an Extra Arguments box
- **added** — External provider has tool calling and tool support
- **added** — Sign in with Codex subscription and use Codex tools inside Chat
- **added** — Paste a YouTube link to attach its transcript including title, channel, duration, link, and caption language
- **added** — Save a full chat or reply into project sources while keeping reasoning, tool calls, and citations
- **added** — Settings > Debugging now shows main server and model-runner logs with live, 3-second, and manual refresh modes
- **added** — Idle image and video models can optionally unload to free VRAM for Chat or Training
- **changed** — Inference is up to 10% faster in supported cases with lower VRAM use and tunable memory limit
- **changed** — MiniMax-H3 can run on smaller supported GPUs by splitting large model parts into pieces that fit
- **changed** — Model picker hides H3 options that current hardware cannot run instead of letting them fail after selection
- **changed** — H3 options are labelled Fast FP8 or Slow to make speed difference clear before downloading
- **changed** — Chat settings stay with the conversation across remote sessions
- **changed** — Chat shows live prompt and generation speeds while long streaming replies use much less CPU
- **changed** — Macs now choose context size from memory that is actually free
- **fixed** — Inference speed for MiniMax-H3 is 10x faster with FP8 quantization
- **fixed** — Bypass permissions fixed
- **fixed** — GGUF image detection when choosing a model for the API
Qwen3.8-27B and Qwen3.8-2.4T can now be run locally in Unsloth!
Run on 17GB RAM via Unsloth Dynamic GGUFs. You can also fine-tune Qwen3.8-27B in Unsloth.
Qwen3.8-27B is by far the strongest model for its size. We also uploaded NVFP4 quants.
Guide: https://unsloth.ai/docs/models/qwen3.8
GGUF: https://huggingface.co/unsloth/Qwen3.8-27B-GGUF
See 1-bit Qwen3.8-2.4T GGUF running in Unsloth:
**[🦥 Download Unsloth Desktop for Linux, Windows, MacOS](https://unsloth.ai/download)**
Here's what you can do with Unsloth Desktop:
- Get **up to 50% more accurate tool calling** with self-healing calls and sandboxed code execution.
- Run **Muse Glimmer 30B, Kimi K3, Qwen3.8, DeepSeek-V4 Flash 0731, Gemma 4**, and more.
- Generate videos with **MiniMax-H3**, and create images and videos with other diffusion models at up to **2× faster inference** on supported workflows.
- Use unlimited private **web search, Deep Research, RAG and MCP**.
- Export models to **NVFP4, GGUF** and other formats.
- Access Unsloth remotely through **Cloudflare HTTPS**.
- Run on **CPU or multiple GPUs** across **NVIDIA, AMD, Intel and Mac**.
- Train models without code, using less time and VRAM.
- Use local models through Unsloth's **OpenAI-compatible API**, or connect **OpenAI and Anthropic** models.
##### Tools, private research + APIs
[Self-healing tool calling](https://unsloth.ai/docs/new/studio/chat#auto-healing-tool-calling) repairs malformed calls instead of dropping them. Models can run Python and Bash inside sandboxed environments, so they can test code, create files and verify their work.
Use unlimited private web search, let [**Deep Research**](https://unsloth.ai/docs/new/studio/chat) plan and produce cited reports, or bring your own files into **RAG**. You can also connect MCP tools for workflows that need external apps, data or actions.
Local models can be served through Unsloth's [**OpenAI-compatible API**](https://unsloth.ai/docs/basics/api) for agents and other clients. Inside Desktop, you can also connect OpenAI and Anthropic as cloud model providers.
##### Muse Glimmer 30B + latest models
Run [**Muse Glimmer 30B**](https://unsloth.ai/docs/models/muse-glimmer) locally for chat, agents, tools and APIs, alongside **Kimi K3, Qwen3.8, DeepSeek-V4 Flash 0731 and Gemma 4**. Download and manage them in one place through Unsloth Desktop.
##### MiniMax-H3 + image and video diffusion
Run [**MiniMax-H3**](https://huggingface.co/unsloth/MiniMax-H3-GGUF) locally for video generation. Create images and videos locally, edit existing images and train supported diffusion models. Use LoRAs, reference images and ControlNet where available, with up to **2× faster inference** on supported workflows.
##### No-code training, export + remote deployment
Pick a model and dataset, adjust the settings and start training. You can train supported LLMs, diffusion models, TTS models and embedding models without writing code. On supported LLM workloads, training is up to 2× faster and uses up to 70% less VRAM.
[Export](https://unsloth.ai/docs/new/unsloth-studio/export) your trained models to **NVFP4, GGUF** and other supported formats. You can also securely deploy and access models remotely: turn on Remote access to publish Unsloth through a Cloudflare HTTPS link, then use the app and its local APIs from another device.
##### Hardware + platform support
Unsloth Desktop runs on **Windows, macOS and Linux**. Hardware support spans **CPU and multi-GPU systems**, **NVIDIA and AMD GPUs**, **Intel hardware**, and **Mac**.
CPU support includes Chat and
_[Truncated at 4000 characters — full notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.702-beta]_
### v0.1.701-beta — Introducing Unsloth Desktop 🦥
- Date: 2026-08-11
- Version: v0.1.701-beta
- Original notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.701-beta
- Permalink: https://whatsnew.fyi/product/unsloth/releases/v0.1.701-beta
- Labels: Pre-release
- **added** — Unsloth Desktop application for running and training AI models locally on Windows, macOS and Linux
- **added** — Self-healing tool calling that repairs malformed calls and executes sandboxed code
- **added** — Support for running Muse Glimmer 30B, Kimi K3, Qwen3.8, DeepSeek-V4 Flash 0731, and Gemma 4 models
- **added** — MiniMax-H3 video generation and image/video diffusion model support
- **added** — Unlimited private web search, Deep Research, RAG and MCP support
- **added** — Model export to NVFP4, GGUF and other formats
- **added** — Remote access through Cloudflare HTTPS for accessing models from other devices
- **added** — OpenAI-compatible API for serving local models
- **added** — No-code training interface without writing code
- **added** — Support for CPU and multiple GPU systems including NVIDIA, AMD, Intel and Mac
- **added** — Image and video editing with LoRAs, reference images and ControlNet support
- **changed** — Training up to 2× faster and using up to 70% less VRAM on supported LLM workloads
- **changed** — Image and video inference up to 2× faster on supported workflows
- **fixed** — Slow Windows downloading by fixing throttling issue, achieving 200× faster speeds
- **fixed** — Mac command line tools download requirement caused by uv bug
- **fixed** — AMD Strix Halo GPU detection
[Unsloth Desktop](https://unsloth.ai/) is here! The first desktop app to run and train AI models locally. Research, export and deploy from the same open-source app on Windows, macOS and Linux.
**[🦥 Download Unsloth Desktop for Linux, Windows, MacOS](https://unsloth.ai/download)**
Here's what you can do with Unsloth Desktop:
- Get **up to 50% more accurate tool calling** with self-healing calls and sandboxed code execution.
- Run **Muse Glimmer 30B, Kimi K3, Qwen3.8, DeepSeek-V4 Flash 0731, Gemma 4**, and more.
- Generate videos with **MiniMax-H3**, and create images and videos with other diffusion models at up to **2× faster inference** on supported workflows.
- Use unlimited private **web search, Deep Research, RAG and MCP**.
- Export models to **NVFP4, GGUF** and other formats.
- Access Unsloth remotely through **Cloudflare HTTPS**.
- Run on **CPU or multiple GPUs** across **NVIDIA, AMD, Intel and Mac**.
- Train models without code, using less time and VRAM.
- Use local models through Unsloth's **OpenAI-compatible API**, or connect **OpenAI and Anthropic** models.
##### Tools, private research + APIs
[Self-healing tool calling](https://unsloth.ai/docs/new/studio/chat#auto-healing-tool-calling) repairs malformed calls instead of dropping them. Models can run Python and Bash inside sandboxed environments, so they can test code, create files and verify their work.
Use unlimited private web search, let [**Deep Research**](https://unsloth.ai/docs/new/studio/chat) plan and produce cited reports, or bring your own files into **RAG**. You can also connect MCP tools for workflows that need external apps, data or actions.
Local models can be served through Unsloth's [**OpenAI-compatible API**](https://unsloth.ai/docs/basics/api) for agents and other clients. Inside Desktop, you can also connect OpenAI and Anthropic as cloud model providers.
##### Muse Glimmer 30B + latest models
Run [**Muse Glimmer 30B**](https://unsloth.ai/docs/models/muse-glimmer) locally for chat, agents, tools and APIs, alongside **Kimi K3, Qwen3.8, DeepSeek-V4 Flash 0731 and Gemma 4**. Download and manage them in one place through Unsloth Desktop.
##### MiniMax-H3 + image and video diffusion
Run [**MiniMax-H3**](https://huggingface.co/unsloth/MiniMax-H3-GGUF) locally for video generation. Create images and videos locally, edit existing images and train supported diffusion models. Use LoRAs, reference images and ControlNet where available, with up to **2× faster inference** on supported workflows.
##### No-code training, export + remote deployment
Pick a model and dataset, adjust the settings and start training. You can train supported LLMs, diffusion models, TTS models and embedding models without writing code. On supported LLM workloads, training is up to 2× faster and uses up to 70% less VRAM.
[Export](https://unsloth.ai/docs/new/unsloth-studio/export) your trained models to **NVFP4, GGUF** and other supported formats. You can also securely deploy and access models remotely: turn on Remote access to publish Unsloth through a Cloudflare HTTPS link, then use the app and its local APIs from another device.
##### Hardware + platform support
Unsloth Desktop runs on **Windows, macOS and Linux**. Hardware support spans **CPU and multi-GPU systems**, **NVIDIA and AMD GPUs**, **Intel hardware**, and **Mac**.
CPU support includes Chat and Data Recipes. Training and inference options vary by model and backend.
##### Download Unsloth Desktop
Unsloth Desktop is **free and open source**. Download it for:
- **Windows**
- **macOS**
- **Linux**
**[🦥 Download Unsloth Desktop](https://unsloth.ai/download)**
**[Star Unsloth on GitHub](https://github.com/unslothai/unsloth)**
##### v0.1.701-beta Update (1PM PDT time August 11th)
- Fixed slow Windows do
_[Truncated at 4000 characters — full notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.701-beta]_
### v0.1.70-beta — Introducing Unsloth Desktop 🦥
- Date: 2026-08-11
- Version: v0.1.70-beta
- Original notes: https://github.com/unslothai/unsloth/releases/tag/v0.1.70-beta
- Permalink: https://whatsnew.fyi/product/unsloth/releases/v0.1.70-beta
- Labels: Pre-release
- **added** — Unsloth Desktop app for running and training AI models locally on Windows, macOS and Linux
- **added** — Self-healing tool calling that repairs malformed calls and executes sandboxed code
- **added** — Support for running Muse Glimmer 30B, Kimi K3, Qwen3.8, DeepSeek-V4 Flash 0731, and Gemma 4 models
- **added** — MiniMax-H3 for local video generation with up to 2× faster inference on supported workflows
- **added** — Private web search, Deep Research, RAG and MCP capabilities
- **added** — Model export to NVFP4, GGUF and other formats
- **added** — Remote access through Cloudflare HTTPS for secure deployment
- **added** — No-code training interface for LLMs, diffusion models, TTS models and embedding models
- **added** — OpenAI-compatible API for serving local models to agents and other clients
- **added** — Support for connecting OpenAI and Anthropic cloud model providers
- **added** — Training up to 2× faster and using up to 70% less VRAM on supported LLM workloads
- **added** — Support for CPU and multi-GPU systems across NVIDIA, AMD, Intel and Mac hardware
[Unsloth Desktop](https://unsloth.ai/) is here! The first desktop app to run and train AI models locally. Research, export and deploy from the same open-source app on Windows, macOS and Linux.
|