Written by Jakub Rusinowski · Last updated July 21, 2026
Ranked for coding and software engineering: Writing, refactoring and debugging code in an editor or terminal, with the model reading real project files.
Best overall: Intel Arc B580
12 GB VRAM at 456 GB/s. It runs 56 of the models that qualify for this workload; the strongest is Qwen 3 14B at an estimated 33 tokens/sec.
| GPU | VRAM | MSRP | Models that fit | Best model it runs | Est. speed | |
|---|---|---|---|---|---|---|
| Best overall | Intel Arc B580 | 12 GB | $249 | 56 | Qwen 3 14B | ~33 tok/s |
| Best value | Intel Arc B580 | 12 GB | $249 | 56 | Qwen 3 14B | ~33 tok/s |
| Budget pick | Intel Arc B570 | 10 GB | $219 | 55 | Qwen3-Coder 8B | ~47.2 tok/s |
| Most memory | NVIDIA DGX Spark | 128 GB | $4,699 | 91 | Devstral-2 123B | ~7.9 tok/s |
| GPU | Score | VRAM | MSRP | Models fit | Est. speed | Tok/s per watt | Cost per model |
|---|---|---|---|---|---|---|---|
| Intel Arc B580 | 72.5 | 12 GB | $249 | 56 | ~33 | 0.17 | $4 |
| NVIDIA GeForce RTX 5090 | 70.2 | 32 GB | $1,999 | 78 | ~57.8 | 0.1 | $26 |
| AMD Radeon RX 7900 XT | 69.9 | 20 GB | $899 | 68 | ~28.6 | 0.09 | $13 |
| AMD Radeon RX 7800 XT | 69.1 | 16 GB | $499 | 63 | ~43.9 | 0.17 | $8 |
| AMD Radeon RX 9070 | 68.8 | 16 GB | $549 | 63 | ~44.9 | 0.2 | $9 |
| NVIDIA GeForce RTX 5070 | 68.4 | 12 GB | $549 | 56 | ~46.9 | 0.19 | $10 |
| AMD Radeon RX 9070 XT | 68.1 | 16 GB | $599 | 63 | ~49.7 | 0.23 | $10 |
| AMD Radeon RX 7900 XTX | 66.5 | 24 GB | $999 | 78 | ~32.9 | 0.09 | $13 |
| NVIDIA GeForce RTX 4090 | 65.9 | 24 GB | $1,599 | 78 | ~34.5 | 0.08 | $21 |
| NVIDIA GeForce RTX 5070 Ti | 65.6 | 16 GB | $749 | 63 | ~60.5 | 0.2 | $12 |
| NVIDIA GeForce RTX 4070 Ti Super | 65.3 | 16 GB | $799 | 63 | ~46.9 | 0.16 | $13 |
| NVIDIA GeForce RTX 4070 | 64.6 | 12 GB | $599 | 56 | ~36.2 | 0.18 | $11 |
The Intel Arc B580 — 12 GB of VRAM runs 56 qualifying models, the strongest being Qwen 3.
The Intel Arc B570 at $219, which runs 55 qualifying models.
8 GB is the entry point at which a model for this workload will run at all. More memory buys a stronger model, not just a faster one.