Please check out our benchmarking guide for benchmarking different models with different backends: DGX Spark Performance FAQ
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| DGX Spark + Qwen3-Next-80B: Proven Performance, But Missing Clear Path to NIM, TensorRT-LLM & Web UIs | 16 | 4053 | March 6, 2026 | |
| DGX Spark performance | 50 | 4442 | February 27, 2026 | |
| Dgx spark benchmark performance | 17 | 1993 | January 4, 2026 | |
| TensorRT-LLM + nvidia/Llama-3.3-70B-Instruct-NVFP4 = 5 tok/s | 4 | 602 | February 1, 2026 | |
| NVIDIA folks -- where is this promised nvfp4 speedup? | 27 | 2576 | March 26, 2026 | |
| Question on Inference Performance Results of Qwen3 235B A22B on 2× DGX Spark | 5 | 709 | December 19, 2025 | |
| We unlocked NVFP4 on the DGX Spark: 20% faster than AWQ! | 145 | 7012 | March 28, 2026 | |
| 6x Spark setup | 112 | 8606 | April 25, 2026 | |
| Qwen3.5-122B-A10B NVFP4 Quantized for DGX Spark — 234GB → 75GB, Runs on 128GB | 44 | 9465 | April 9, 2026 | |
| DGX Spark: The Sovereign AI Stack — Dual-Model Architecture for Local Inference | 9 | 1662 | February 13, 2026 |