# Best GPU for local AI by budget | What Can I Run AI?

> The best GPUs, Macs and mini PCs for running local LLMs under $300, $500, $1000, $2000 and more, ranked by usable memory.

Source: https://whatcanirunai.vercel.app/best-gpu-under/

# Best GPU for local AI by budget

Pick your budget. Each page ranks every GPU, Mac and mini PC we track at or under that price by usable memory, with price per GB and which models fit.

## Budgets

__Top pick at each budget__
| Budget                                                   | Most memory for the money                                                             | Usable | Biggest model at Q4               |
| -------------------------------------------------------- | ------------------------------------------------------------------------------------- | ------ | --------------------------------- |
| [Best GPU under $300](https://whatcanirunai.vercel.app/best-gpu-under/300/)8 devices     | [Intel Arc B580](https://whatcanirunai.vercel.app/hardware/arc-b580/)$249                                             | 12 GB  | OLMo-2-1124-13B-Instruct          |
| [Best GPU under $500](https://whatcanirunai.vercel.app/best-gpu-under/500/)25 devices    | [Radeon RX 7600 XT](https://whatcanirunai.vercel.app/hardware/rx-7600-xt/)$329                                        | 16 GB  | Ling-Coder-lite                   |
| [Best GPU under $800](https://whatcanirunai.vercel.app/best-gpu-under/800/)44 devices    | [Radeon RX 7600 XT](https://whatcanirunai.vercel.app/hardware/rx-7600-xt/)$329                                        | 16 GB  | Ling-Coder-lite                   |
| [Best GPU under $1,000](https://whatcanirunai.vercel.app/best-gpu-under/1000/)53 devices | [Intel Arc Pro B70 32GB](https://whatcanirunai.vercel.app/hardware/arc-pro-b70/)$949                                  | 32 GB  | IQuest-Coder-V1-40B-Loop-Instruct |
| [Best GPU under $1,500](https://whatcanirunai.vercel.app/best-gpu-under/1500/)59 devices | [Intel Arc Pro B70 32GB](https://whatcanirunai.vercel.app/hardware/arc-pro-b70/)$949                                  | 32 GB  | IQuest-Coder-V1-40B-Loop-Instruct |
| [Best GPU under $2,000](https://whatcanirunai.vercel.app/best-gpu-under/2000/)63 devices | [Intel Arc Pro B70 32GB](https://whatcanirunai.vercel.app/hardware/arc-pro-b70/)$949                                  | 32 GB  | IQuest-Coder-V1-40B-Loop-Instruct |
| [Best GPU under $3,500](https://whatcanirunai.vercel.app/best-gpu-under/3500/)68 devices | [Framework Desktop (Ryzen AI Max+ 395) 128GB](https://whatcanirunai.vercel.app/hardware/framework-desktop-128/)$3,449 | 96 GB  | Mistral-Small-4-119B-2603         |

Prices are dated snapshots from our data (GPUs at official list price / MSRP). Not sure how much memory you need? See [best LLMs by VRAM size](https://whatcanirunai.vercel.app/best-llm-for/) or the [VRAM calculator](https://whatcanirunai.vercel.app/vram-calculator/).

## Quick answers

What is the cheapest way to run local LLMs?

The cheapest device we track is the Intel Arc B570 at $219 with 10 GB, enough for 340 of our 656 models at Q4.

How much should I spend on a GPU for local AI?

Memory matters most. Under $300 you get 12 GB at most; at $3,500 you can reach 96 GB usable. Pick the budget page that matches yours to see what each step up unlocks.

[Buying guides by model family](https://whatcanirunai.vercel.app/best-gpu-for/) · [GPU vs GPU](https://whatcanirunai.vercel.app/compare/)

---
Agent guide: https://whatcanirunai.vercel.app/llms.txt · JSON API: https://whatcanirunai.vercel.app/developers/