Flopper.io GPU Comparison Sheet
NVIDIA GB200 Grace Blackwell Superchip 372GB vs NVIDIA Jetson AGX Thor T5000 128GB
Side by side datacenter GPU specifications from the Flopper.io database.
Overview
| NVIDIA GB200 Grace Blackwell Superchip 372GB | NVIDIA Jetson AGX Thor T5000 128GB | |
|---|---|---|
| FP32 TFLOPS | 160.0 | 8.1 |
| Memory | 372 GB | 128 GB |
| Memory Bandwidth | 16.0 TB/s | 273 GB/s |
| TDP | — | 130 W |
Performance by Precision
| Precision | NVIDIA GB200 Grace Blackwell Superchip 372GB peak TFLOPS (dense) | NVIDIA Jetson AGX Thor T5000 128GB peak TFLOPS (dense) |
|---|---|---|
| FP64 | 80.0 | — |
| FP32 | 160.0 | 8.1 |
| TF32 | 2500.0 | — |
| BF16 | 5000.0 | — |
| FP16 | 5000.0 | — |
| FP8 | 10000.0 | 517.0 |
| FP6 | 10000.0 | — |
| FP4 | 20000.0 | 1035.0 |
| INT8 | 10000.0 | 517.0 |
Specifications
| NVIDIA GB200 Grace Blackwell Superchip 372GB | NVIDIA Jetson AGX Thor T5000 128GB | |
|---|---|---|
| Architecture | Blackwell | Blackwell |
| Process Node | 4nm | — |
| Launch Year | 2024 | 2025 |
| Form Factor | Superchip | Module |
| Memory Type | HBM3e | LPDDR5X |
| Interface Width | — | 256-bit |
| Interconnect | NVLink-C2C | — |
| TDP | — | 130 W |
| FP32 Efficiency | — | 0.06 TFLOPS/W |