Leaderboard Β· All comparisons Β· LynxBenchAI 26Q3 Index

DGX Spark (NVIDIA GB10) vs NVIDIA GeForce RTX 4060 for AI

Device pages: DGX Spark (NVIDIA GB10) Β· NVIDIA GeForce RTX 4060

On the LynxBenchAI 26Q3 Index, the DGX Spark (NVIDIA GB10) scores 1.13x the NVIDIA GeForce RTX 4060 overall (Training 1.12x, Inference 1.18x, Compute 1.03x).

Scores on the LynxBenchAI 26Q3 Index

Ratio = the DGX Spark (NVIDIA GB10)'s score divided by the RTX 4060's
ScoreRatioDGX Spark (NVIDIA GB10)RTX 4060
Overall (GT)1.13x119.69106.26
Training1.12x92.1681.93
Inference1.18x182.32154.96
Compute1.03x84.6782.26

When to Choose Which

  • Choose the DGX Spark (NVIDIA GB10) for training (1.12x the RTX 4060).
  • Choose the DGX Spark (NVIDIA GB10) for inference (1.18x the RTX 4060).
  • Level on raw compute: the two are within 10% (1.03x).

From the LynxBenchAI 26Q3 Index category scores, which are not raw throughput; a lead under 10% counts as level.

Per Model and Precision

LynxBenchAI score per model and precision; ratio = DGX Spark (NVIDIA GB10) / RTX 4060
ModelPrecisionRatioDGX Spark (NVIDIA GB10)RTX 4060
compute
Dense MatMulBF162.93x176.1660.09
Dense MatMulFP162.91x173.4059.51
Dense MatMulFP323.07x154.7450.46
Dense MatMulFP641.86x3.011.62
Sparse MatMulFP320.40x62.65155.55
Sparse MatMulFP640.33x48.51146.47
inference
DistilBERTFP161.53x347.13226.96
DistilBERTFP81.30x185.12142.44
DistilBERTINT81.04x141.54136.64
DistilGPT2FP161.50x300.91200.54
DistilGPT2FP81.23x153.69125.43
DistilGPT2INT81.04x133.00128.30
Flan-T5 SmallFP161.53x467.15304.88
Flan-T5 SmallFP81.18x242.60205.70
Flan-T5 SmallINT80.86x146.05169.75
GraphGPS (Peptides)FP160.97x89.6892.03
GraphGPS (Peptides)FP80.93x43.9647.11
GraphGPS (Peptides)INT80.88x40.7846.56
MobileNetV3FP161.11x407.57365.84
MobileNetV3INT80.90x90.32100.37
training
DistilBERTBF161.33x140.71105.77
DistilBERTFP321.59x210.57132.29
DistilGPT2BF161.14x72.4263.43
DistilGPT2FP321.49x146.3898.36
Flan-T5 SmallBF160.84x54.3964.74
Flan-T5 SmallFP320.89x107.25120.25
GraphGPS (Peptides)BF161.12x45.7440.90
GraphGPS (Peptides)FP321.08x43.0839.80
MobileNetV3BF160.97x82.1984.66
MobileNetV3FP320.89x100.98113.22

Related Comparisons

Each figure is the first device's overall score divided by the second's.

Measure your own machine: pip install lynxbench-ai Β· Install guide β†’

How to Read These Results

A LynxBenchAI score is sustained throughput on a fixed catalogue of PyTorch models, measured after a warm-up at a batch size raised until throughput stops improving, so it is what the machine holds under load, not its best instant. Training, Inference and Compute are three separate scores; GT combines them (Training and Inference weigh 0.4 each, Compute 0.2).

Every run uses the stock, publicly released PyTorch stack, so a result is broadly indicative of the out-of-the-box PyTorch experience on that hardware, and so of custom PyTorch models too. It does not run your own model, and it is not a vendor-tuned peak like an MLPerf submission.

Scores compare only within one Index: the 26Q3 Index fixes the models, precisions, correctness thresholds and scoring, and a later Index may change them. Each test is one timed window, the scores have no 0-100 scale, and a precision a device cannot run scores zero rather than being skipped.

LynxBenchAI is built by TechnoLynx, which the European Commission engaged as an independent expert advisor on AI performance and benchmarking for AI Gigafactories under the EuroHPC Joint Undertaking (case study).

Questions

Is the DGX Spark (NVIDIA GB10) faster than the NVIDIA GeForce RTX 4060 for AI?
On the LynxBenchAI 26Q3 Index, the DGX Spark (NVIDIA GB10) scores 1.13x the NVIDIA GeForce RTX 4060 overall (Training 1.12x, Inference 1.18x, Compute 1.03x).
Should I choose the DGX Spark (NVIDIA GB10) or the NVIDIA GeForce RTX 4060?
Choose the DGX Spark (NVIDIA GB10) for training (1.12x the RTX 4060). Choose the DGX Spark (NVIDIA GB10) for inference (1.18x the RTX 4060). Level on raw compute: the two are within 10% (1.03x). These follow the LynxBenchAI 26Q3 Index category scores, which are not raw throughput.
Where is the difference between the DGX Spark (NVIDIA GB10) and the NVIDIA GeForce RTX 4060 largest?
On the LynxBenchAI 26Q3 Index, the ratio is highest on Dense MatMul at FP32 (Compute), where the DGX Spark (NVIDIA GB10) scores 3.07x the NVIDIA GeForce RTX 4060, and lowest on Sparse MatMul at FP64 (Compute), at 0.33x.
What does this result tell me about my own PyTorch model?
Every run uses the stock, publicly released PyTorch stack, so a result is broadly indicative of the out-of-the-box PyTorch experience on that hardware, and so of custom PyTorch models too. It does not run your own model, and it is not a vendor-tuned peak like an MLPerf submission.

The Personal Edition submits your benchmark results to TechnoLynx's servers as part of participating in the public leaderboard. If that isn't acceptable for your use case, contact TechnoLynx about Press, Pro, or Enterprise licenses.

Get Started