NVIDIA DGX Spark vs AMD Radeon Pro V620 32GB for local LLMs
94 reports on the NVIDIA DGX Spark and 12 on the AMD Radeon Pro V620 32GB
Which is faster for local LLMs?
Nobody has reported the same model at the same quant on both cards yet, so there's no measured head-to-head. Memory bandwidth, which caps how fast a card can write, is 1.9x higher on the AMD Radeon Pro V620 32GB (273 GB/s on the NVIDIA DGX Spark, 512 on the AMD Radeon Pro V620 32GB). The estimates below compare them on the models people run most.
- Memory
- 128GB unified
- Memory bandwidth
- 273 GB/s
- FP16 compute
- 125 TFLOPS
- Reports
- 94
- Memory
- 32GB
- Memory bandwidth
- 512 GB/s
- FP16 compute
- 41.5 TFLOPS
- Reports
- 12
Estimated by the calculator
The quant the calculator recommends for each card and its estimated speed at a 32,768-token context with 32 GB of system RAM, from memory bandwidth and the model's shape.
| Model | NVIDIA DGX Spark | AMD Radeon Pro V620 32GB |
|---|---|---|
| Qwen3.8 27B | 4.0 t/sFP16 | 12.8 t/sQ6_K |
| Qwen3.8 125B · 6B active | 14.4 t/sQ6_K | 18.1 t/sQ2_K, experts in RAM |
| Qwen3.6 35B · 3B active | 14.5 t/sFP16 | 46.6 t/sQ5_K_M |
| DeepSeek V4 Flash 284B · 13B active | 38.3 t/sQ2_K | doesn't fit at these settings |
| Qwen3.6 27B | 4.0 t/sFP16 | 12.8 t/sQ6_K |
| Gemma 4 26B · 4B active | 10.8 t/sFP16 | 32.7 t/sQ6_K |
Change the context, RAM or model in the calculator: NVIDIA DGX Spark · AMD Radeon Pro V620 32GB
Every report on each card, with its source: NVIDIA DGX Spark · AMD Radeon Pro V620 32GB