Flopper.io GPU Comparison Sheet
October 20, 2026
flopper.io/compare/nvidia-gb300-nvl72-gpu-288gb-vs-openai-jalapenoNVIDIA GB300 NVL72 GPU 288GB vs OpenAI Jalapeño
Side by side datacenter GPU specifications from the Flopper.io database.
Overview
| NVIDIA GB300 NVL72 GPU 288GB | OpenAI Jalapeño | |
|---|---|---|
| FP32 TFLOPS | 83.3 | — |
| Memory | 288 GB | 216 GB |
| Memory Bandwidth | 8.0 TB/s | 15.4 TB/s |
| TDP | — | 700 W |
Performance by Precision
| Precision | NVIDIA GB300 NVL72 GPU 288GB peak TFLOPS (dense) | OpenAI Jalapeño peak TFLOPS (dense) |
|---|---|---|
| FP64 | 1.4 | — |
| FP32 | 83.3 | — |
| TF32 | 1250.0 | — |
| BF16 | 2500.0 | — |
| FP16 | 2500.0 | — |
| FP8 | 5000.0 | 3400.0 |
| FP6 | 5000.0 | — |
| FP4 | 15000.0 | 13400.0 |
| INT8 | 165.0 | — |
Specifications
| NVIDIA GB300 NVL72 GPU 288GB | OpenAI Jalapeño | |
|---|---|---|
| Architecture | Blackwell Ultra | Jalapeño |
| Process Node | 4nm | — |
| Launch Year | 2025 | 2026 |
| Form Factor | SXM | Custom |
| Memory Type | HBM3e | HBM4 |
| Interface Width | — | — |
| Interconnect | NVLink | Ethernet scale-up (Broadcom Tomahawk 6) |
| TDP | — | 700 W |
| FP32 Efficiency | — | — |