Best graphics cards for local AI under $700
Ranked by what each one runs fully in memory at 32k context, with prices in Canada today.
In short
The top pick in Canada right now is the Intel Arc A770 16GB, $655 at Amazon: it runs gpt-oss 20B fully in memory at 32k context at about 48 tok/s (estimated). Ranked by the most capable open model each one runs fully in memory at 32k context and a usable speed, then by how many models it runs, then by speed on Qwen 3.8 27B, then by price.
Updated . Speeds are estimates; how they are worked out.
The list
Best first. Speeds are estimates est.
| # | Card | Memory | Price | Best model it runs at 32k | Models in memory | Qwen 3.8 27B | tok/s per $1,000 |
|---|---|---|---|---|---|---|---|
| 1 | Intel Arc A770 16GB | 16 GB | $655Amazon | gpt-oss 20BMXFP448 tok/s | 13 | offload | – |
| 2 | NVIDIA GeForce RTX 3060 12GB | 12 GB | $497Amazon | Gemma 4 12BQ6_K21 tok/s | 10 | offload | – |
| 3 | Intel Arc B580 | 12 GB | $560Amazon | Gemma 4 12BQ6_K21 tok/s | 10 | offload | – |
Questions
- What is the best graphics card for running AI models locally under $700?
- The top pick in Canada right now is the Intel Arc A770 16GB, $655 at Amazon: it runs gpt-oss 20B fully in memory at 32k context at about 48 tok/s (estimated).