GH200 (96 GB) specs
NVIDIA · Shipping now · first available Nov 2023. Dense compute, memory, bandwidth and power from the vendor's documents; each precision stays separate.
All accelerators · Compare with H100 NVL
Specifications
| Dense BF16 · FP32 accumulate | 990 TFLOPS |
|---|---|
| Dense FP8 · FP32 accumulate | 1,979 TFLOPS |
| Dense FP4 · FP32 accumulate | Not published |
| Memory | 96 GB · HBM |
| Memory bandwidth | 4 TB/s |
| Maximum chip power | Not published |
| Interconnect | NVLink-C2C · 900 GB/s |
| First availability | Nov 2023 |
| Availability | Shipping |
The Hopper GPU of a Grace Hopper superchip, in its 96 GB HBM3 version (the one rented by the hour). NVIDIA publishes power only for the whole module, CPU and memory included.
Dense tensor rates with FP32 accumulation, never sparse or FP16-accumulate headline figures. Memory is in decimal GB. The est marker flags a value its vendor-document review did not confirm.
Rent it
No tracked rental quote. The site tracks providers' published list prices; this part may still be rentable under terms it does not record.
Value ratios
| $ per dense BF16 PFLOP-hour | No on-demand quote |
|---|---|
| Dense BF16 TFLOPS per watt | No power figure |
| Bandwidth bytes per 1,000 BF16 FLOP | 4.04 |
| $ per dense FP8 PFLOP-hour | No on-demand quote |
| Dense FP8 TFLOPS per watt | No power figure |
| Bandwidth bytes per 1,000 FP8 FLOP | 2.02 |
| $ per GB of memory per hour | No on-demand quote |
Prices divide the cheapest on-demand list price; theoretical ratios do not measure application speed.
Sources
Specifications reviewed . 96 GB HBM3 column: 990 BF16 and 1,979 FP8 TFLOPS dense, up to 4 TB/s, NVLink-C2C 900 GB/s. Power is published only for the module (450 to 1,000 W, memory, CPU and GPU together), so power stays empty.
First availability: Lambda Cloud Clusters with GH200 ↗ (2023-11-13): “Lambda Cloud Clusters are now available with the NVIDIA GH200 Grace Hopper Superchip”