NVIDIA · GPU · Results: 26Q3, recorded 2026-08-26
| Model | Precision | Sustained |
|---|---|---|
| MobileNetV3 | FP32 | 177.03 |
| MobileNetV3 | BF16 | 122.02 |
| Flan-T5 Small | FP32 | 192.87 |
| Flan-T5 Small | BF16 | 108.99 |
| DistilGPT2 | FP32 | 171.39 |
| DistilGPT2 | BF16 | 93.30 |
| DistilBERT | FP32 | 206.40 |
| DistilBERT | BF16 | 168.12 |
| GraphGPS (Peptides) | FP32 | 70.84 |
| GraphGPS (Peptides) | BF16 | 76.01 |
| Model | Precision | Sustained |
|---|---|---|
| MobileNetV3 | FP16 | 629.11 |
| MobileNetV3 | INT8 | 175.89 |
| Flan-T5 Small | FP16 | 506.41 |
| Flan-T5 Small | INT8 | 275.77 |
| Flan-T5 Small | FP8 | 317.32 |
| DistilGPT2 | FP16 | 276.66 |
| DistilGPT2 | INT8 | 154.57 |
| DistilGPT2 | FP8 | 156.69 |
| DistilBERT | FP16 | 356.37 |
| DistilBERT | INT8 | 220.84 |
| DistilBERT | FP8 | 227.87 |
| GraphGPS (Peptides) | FP16 | 138.61 |
| GraphGPS (Peptides) | INT8 | 67.70 |
| GraphGPS (Peptides) | FP8 | 70.83 |
| Model | Precision | Sustained |
|---|---|---|
| Dense MatMul | FP32 | 93.06 |
| Dense MatMul | FP16 | 96.21 |
| Dense MatMul | BF16 | 97.06 |
| Dense MatMul | FP64 | 2.48 |
| Sparse MatMul | FP32 | 60.39 |
| Sparse MatMul | FP64 | 44.66 |
Results: 26Q3 · GT = LynxBenchAI Global Score. Every precision result is sustained throughput under a declared, bounded optimisation budget — no collapsed single number below the category level.
← Back to the leaderboardThe 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.
LynxBenchAI End-User License Agreement · LynxBenchAI Privacy Policy