24 GB
35 models are a fit or a tight fit on the smallest pack, at 24 GB of memory with no Mac or CPU discount.
Qwen2.5-7B-Instruct, Qwen3-8B, Qwen3.5-4B, Qwen3.5-9B, Qwen3.8-27B, Qwen2.5-VL-7B-Instruct and 29 more
Compute
Rent memory when the model does not fit.
A rental is someone else's GPU for the time you need it. AI Price Desk does not quote an hourly rate. The provider's page is the price, and that price moves.
48 GB
38 models are a fit or a tight fit on the smallest pack, at 48 GB of memory with no Mac or CPU discount.
Qwen2.5-7B-Instruct, Qwen3-8B, Qwen3.5-4B, Qwen3.5-9B, Qwen3.8-27B, Qwen2.5-VL-7B-Instruct and 32 more
80 GB
40 models are a fit or a tight fit on the smallest pack, at 80 GB of memory with no Mac or CPU discount.
Qwen2.5-7B-Instruct, Qwen3-8B, Qwen3.5-4B, Qwen3.5-9B, Qwen3.8-27B, Qwen2.5-VL-7B-Instruct and 34 more
141 GB
40 models are a fit or a tight fit on the smallest pack, at 141 GB of memory with no Mac or CPU discount.
Qwen2.5-7B-Instruct, Qwen3-8B, Qwen3.5-4B, Qwen3.5-9B, Qwen3.8-27B, Qwen2.5-VL-7B-Instruct and 34 more
What this page does not do
It does not start a machine, hold an account, or take a payment. It tells you which memory class the fit formula marks as enough. Confirm the runtime on the publisher's repository before you rent, because a number that fits the weights can still miss a framework the host does not offer.
This directory is assembled from public model records. Freshness is age. Popularity is downloads and likes. A fit figure is an estimate from a parameter count or from a published file size, not a run on our hardware. How the numbers are made. AI disclosure.