What we include
Companies that sell LLM fine-tuning as a product (managed fine-tuning APIs, training APIs, GPU platforms used for fine-tuning, and hyperscaler tuning services) or as a service (custom model training for enterprises), with an active public website and an available offering. Products that are closed to new customers or at end of life are left out and noted on vendor profiles where relevant.
Sources
Each fact comes from the vendor’s own website, documentation, pricing page, trust portal, or a primary release such as a company announcement or regulatory filing. Every profile lists its sources and the date we checked them.
What we do not do
We do not test vendors, score them, rank them, or publish reviews. We do not claim partnerships or customer relationships we cannot source. Where a vendor publishes no price, we say “Contact vendor”. Where we found no certification, we say so and suggest asking for documentation; that is not a finding that the vendor lacks one.
Pricing
Where a vendor publishes fine-tuning prices, we quote them exactly as published on the checked date, with the unit (per training token, per GPU-hour, or per hour). Prices change often; confirm on the vendor’s page before you budget.
Certifications
Listed as the vendor states them. For cloud providers, a certification usually covers a list of in-scope services; scope can differ by product, region, or subsidiary, so we recommend requesting reports.
Order
Vendor lists default to alphabetical order. Filters and comparisons change only what is shown, not any ranking.
Updates
We re-check sources at least every 90 days and when vendors announce material changes.
Commercial policy
Today we accept no payment for listings or placement. If sponsorship is introduced later, sponsored items will be labeled, kept out of editorial comparisons and filters’ default order, and disclosed on this page.