Unsloth
Free tierRun and Train Models Locally
Free tier available·All audiences·API available·Open source
Key strengths
Local-first AI model running and trainingOpen-source and free desktop appOpenAI-compatible API for agent integrationSignificant VRAM reduction (up to 90% on Enterprise)Significant training speed improvements (up to 32x on Enterprise)Day Zero support for latest modelsImage and video generation supportSelf-healing tool calls for improved accuracyCross-platform (Mac, Windows, Linux)
Free tier + paid plans
Founded 2023
Self-hostable
No ratings yet
Unsloth — Technical Documentation Highlights
Pricing & Tiers
| Tier | Speed Boost | VRAM Reduction | GPU Scale | Pricing |
|---|---|---|---|---|
| Free (OSS) | 2x | 60% | Single GPU | Free |
| Pro | 2.5x | 80% | Up to 8 GPUs | Contact |
| Enterprise | 32x | Up to 90%+ | Multi-node | Contact |
Fine-Tuning
- Supports 4-bit and 16-bit LoRA fine-tuning out of the box on the free tier
- Enterprise tier adds up to 30% accuracy improvements over baseline fine-tuning
API Surface
- OpenAI-compatible API endpoint for seamless integration with existing LLM tooling, agent frameworks, and SDKs
- Compatible with Claude Code and Codex for AI-assisted development workflows
- Cloudflare tunnel integration for secure remote access to locally hosted models
- GGUF and MLX format support for quantized model deployment
Reliability
- Self-healing tool calls improve robustness of function-calling in agentic and multi-step reasoning pipelines
Deployment
- Self-hostable on-premises or on local workstations
- Cross-platform: Mac, Windows, Linux
- Free open-source desktop app included
- Day Zero model support ensures compatibility with the latest model releases