Unsloth

5x faster, 80% less VRAM fine-tuning for LLMs.

Update, Nov 12, 2026:Now ships Unsloth Desktop, a free open-source app for Mac, Windows and Linux to run, train and serve models locally. The core library is Apache 2.0; the Studio UI is AGPL-3.0.

Rank
#15 of 66#6 of 17 in Infrastructure
Adoption score
84 / 1003.5 points above the category average
Entry price
FreeConfirmed Nov 12, 2026
Open source
YesCan be self-hosted

Our assessment

Unsloth ranks #6 of 17 in Infrastructure and #15 of 66 overall, with an adoption score of 84, 3.5 points above its category average. Built by Unsloth AI, it delivers 5x faster, 80% less VRAM fine-tuning for LLMs.

It is open source under the Apache-2.0 / AGPL-3.0 license, so you can self-host it, read the code and avoid vendor lock-in. Entry pricing is Free, confirmed on the official page.

Where it fits: Ultra-fast parameter-efficient fine-tuning (LoRA/QLoRA) on consumer GPUs.

Prices and facts come from Unsloth's own pages; the full list is on the sources page. Suggest an edit.

Best for
Ultra-fast parameter-efficient fine-tuning (LoRA/QLoRA) on consumer GPUs
Watch out for
You run and maintain it yourself unless a managed tier is offered.
Consider instead
Ollama (Free) · vLLM (Free)

How it compares in Infrastructure

History

Tracked since Oct 13, 2026.

SnapshotRankAdoptionEntry pricePrice check
Nov 20, 2026#15 same84FreeConfirmed
Nov 12, 2026#15 same84FreeConfirmed
Oct 29, 2026#15 same84FreeConfirmed
Oct 15, 2026#15 same84FreeConfirmed
Oct 13, 2026#15 84FreeConfirmed

Key facts

CategoryInfrastructure
MakerUnsloth AI
LicenseApache-2.0 / AGPL-3.0
Open sourceYes
Self-hostYes
Entry priceFree
Category rank#6 of 17
Price checkConfirmed on the vendor's own page

Compare Unsloth

Get an email when the index changes

Price changes, new tools, acquisitions and shutdowns, sent when the data is updated.