Unsloth
Run and fine-tune LLMs and diffusion models locally.
At a glance
Free · Source review 2026-11-02- Price & billing
Open-source under Apache License 2.0 on the observed repository page; free notebooks are also offered.
- Free allowance
Open-source under Apache License 2.0 on the observed repository page; free notebooks are also offered.
- Best suited to
Running and training LLMs and diffusion models locally.
Fit is editorial guidance, not a measured result.- Platforms
Access: not separately verified. Operating systems not separately verified.
Published-source facts and editorial guidance. No product test is implied. Confirm current conditions before purchase.
No product screenshot is claimed for this listing. Use the official website ↗ to inspect the current interface.
What to know before choosing
Good fit
Running and training LLMs and diffusion models locally.
Limits and trade-offs
Local hardware is required; exposing the service needs password care.
Plans and free allowances
Open-source under Apache License 2.0 on the observed repository page; free notebooks are also offered.
Evidence
Official Unsloth GitHub repository ↗ (reviewed 2026-11-02)
Official published page review; no authenticated product test
Dates: Last reviewed 2026-11-02. Last reviewed is the latest source or editorial review date; it is not a content-change date.
A closer look at Unsloth
Unsloth is a local UI to run and train LLMs and diffusion models, the first desktop app to run and train models, supporting GGUF, MLX, Qwen, DeepSeek, Gemma, FLUX and more.
Editorial guidance based on published documentation. The workflow below is a suggested starting point, not a measured product result.
When to choose Unsloth
- A developer wants to run and train LLMs, diffusion, TTS or embedding models locally on Windows, Linux, WSL or macOS.
- A team wants fine-tuning with LoRA, QLoRA, full fine-tuning, RL, GRPO or DPO at 2x speed with 70% less VRAM as claimed.
- A builder wants local models connected to Claude Code, Codex and other agents with one command.
When to choose something else
- A non-technical user wanting a hosted chatbot will find a local training tool and GPU setup more than needed.
- A team needing a hosted commercial plan should note the observed page is an open-source repository, not a SaaS pricing page.
A practical starting workflow
- Download the native Unsloth Desktop app for the operating system, or install manually via the install script, Homebrew, Winget or Docker.
- Run and train models locally: LLMs, MLX, GGUF, diffusion, decision, embedding and audio models, with multi-GPU and NVIDIA, AMD, Intel or CPU support.
- Connect local models to agents with one command, such as unsloth start claude, and serve models through an OpenAI-compatible API.
- Fine-tune with LoRA, QLoRA, full fine-tuning or RL, then export or deploy models in GGUF, NVFP4, FP8 and more formats.
Check before committing
- Local training needs suitable GPU or CPU hardware; speed and VRAM claims are vendor-presented.
- Exposing Unsloth over LAN or HTTPS requires password care, per the repository's own warning.
- This entry describes published repository features and does not compare training quality through testing.
Common questions
What is Unsloth?
The first desktop app to run and train models, with a local UI for LLMs and diffusion models, plus Studio web UI and Core code-based versions.
How do I install it?
Download Unsloth Desktop for Windows, macOS or Linux, or install manually with the install script, Homebrew, Winget or the Docker image unsloth/unsloth.
What fine-tuning does it support?
Reinforcement learning, LoRA, QLoRA, full fine-tuning, pretraining, RL, GRPO, DPO and FP8, claimed at 2x faster with 70% less VRAM and no accuracy loss.
Compare approaches for this task
- UBIAI — Fine-tuning domain-specific LLM components without ML expertise.
- Unsiloed AI — Turning enterprise documents into LLM-ready JSON and Markdown at scale.
Sources and checks
Last source review is not a date of product testing. No rating or test result is implied by inclusion. Check the current plan and terms before buying.