Provider pricing

Unsloth AI API pricing

Unsloth is an open-source library and managed service for fine-tuning open LLMs 2x faster with 70% less VRAM. The hosted Unsloth Cloud runs LoRA, QLoRA, and full fine-tunes on Llama, Mistral, Qwen, and Gemma with per-GPU-hour billing.

Founded2023
HeadquartersSan Francisco, CA
Tracked modelsDirectory

Best for

Researchers and startups fine-tuning open LLMs on tight GPU budgets.

Strengths

  • 2x faster fine-tuning, 70% less GPU memory vs. baseline
  • Open-source library with permissive Apache-2.0 license
  • Supports LoRA, QLoRA, DPO, ORPO, and full fine-tuning
  • Per-GPU-hour pricing on managed cloud