Home › GPUs & Macs › NVIDIA DGX Spark (GB10) vs NVIDIA RTX 4060
NVIDIA DGX Spark (GB10) vs NVIDIA RTX 4060 for local LLMs
Across 4 tests measured on both, the NVIDIA DGX Spark (GB10) is faster in 2 and generates a median 1.49× the tokens/s of the NVIDIA RTX 4060.
On gpt-oss 120B (MXFP4, llama-bench) the NVIDIA DGX Spark (GB10) generates 58.0 tokens/s and the NVIDIA RTX 4060 18.9 tokens/s (measured).
Same model, same quant, same test
| Model | Quant / test | DGX Spark (GB10) tok/s | RTX 4060 tok/s | Faster | Sources and run details |
|---|---|---|---|---|---|
| gpt-oss 120B | MXFP4 llama-bench | 58.0 CUDA | 18.9 CUDA | DGX Spark (GB10) +207% | llama.cpp benches DGX Spark (GB10) · llama.cpp #15396 RTX 4060 |
| gpt-oss 20B | MXFP4 llama-bench | 84.2 CUDA | 42.3 CUDA | DGX Spark (GB10) +99% | llama.cpp benches DGX Spark (GB10) · llama.cpp #15396 RTX 4060 |
| Llama 3.1 8B Instruct | Q4_K_M LocalScore | 34.2 llamafile | 38.3 llamafile | RTX 4060 +12% | LocalScore DGX Spark (GB10) · LocalScore RTX 4060 |
| Llama 3.2 1B Instruct | Q4_K_M LocalScore | 132 llamafile | 134 llamafile | even | LocalScore DGX Spark (GB10) · LocalScore RTX 4060 |
Generation speed in tokens/s, best measured backend for each chip (shown under the number). Backends can differ between the two chips; check them before drawing conclusions. Measured results only, nothing estimated.
Specs
| NVIDIA DGX Spark (GB10) | NVIDIA RTX 4060 | |
|---|---|---|
| Memory | 128 GB | 8 GB |
| Memory bandwidth | 273 GB/s | 272 GB/s |
All results for the NVIDIA DGX Spark (GB10) · All results for the NVIDIA RTX 4060 · More comparisons