Can I deploy NVIDIA H200 NVL or AMD MI250X on Runcrate today?

NVIDIA H200 NVL is on demand in 2 regions from $4.63/hr, billed by the second on most GPUs. AMD MI250X is not in stock on Runcrate right now; leave your email on its pricing page to register interest.

How much VRAM does each GPU have?

NVIDIA H200 NVL has 141GB HBM3e. AMD MI250X has 128GB HBM2e.

H200 NVL or MI250X: which holds a 70B model in 8-bit weights?

It needs 70 GB of 8-bit weights. On the H200 NVL: 1 GPU, 71 GB spare. On the MI250X: 1 GPU, 58 GB spare. Pick the H200 NVL: same GPU count, 46% more memory bandwidth.

GPU comparison

H200 NVL vs MI250X.

Pick the H200 NVL for speed: 2.2x the FP16 throughput and 46% more memory bandwidth. Memory is close (141 GB vs 128 GB).

Launch H200 NVL Reserve a cluster

Rent either on Runcrate

H200 NVLfrom $4.63/hr
$0.033 per GB/hr$0.55 per 100 TFLOPS/hr
Launch H200 NVL H200 NVL pricing
MI250XNot in stock right now
Register interest MI250X pricing

Live hourly price, then per GB of memory and per 100 FP16 TFLOPS.

Which is better for

Which to pick for each job. Memory fit and throughput, from the datasheets.

JobH200 NVLMI250XPick

Hold a 70B model in 8-bit weights
70 GB of 8-bit weights
H200 NVL: 1 GPU, 71 GB spare
MI250X: 1 GPU, 58 GB spare
H200 NVL. Same GPU count, 46% more memory bandwidth.
Serve an 8B model in 16-bit weights
16 GB of 16-bit weights
H200 NVL: 1 GPU, 125 GB spare
MI250X: 1 GPU, 112 GB spare
H200 NVL. Same GPU count, 46% more memory bandwidth.
Fine-tune a 70B model with QLoRA
About 46 GB with 4-bit weights
H200 NVL: 1 GPU, 95 GB spare
MI250X: 1 GPU, 82 GB spare
H200 NVL. Same GPU count, 2.2x the FP16 throughput.
Train on a full node
FP16 tensor throughput per node
H200 NVL: 8x, 6,680 TFLOPS, NVLink in groups of 4
MI250X: 8x, 3,064 TFLOPS
H200 NVL. 2.2x the FP16 throughput per GPU.
Throughput per watt
FP16 TFLOPS per 100 W of TDP
H200 NVL: 139 TFLOPS per 100 W
MI250X: 68 TFLOPS per 100 W
H200 NVL. 2.0x the FP16 per watt.

Compare another pair

Hopper

NVIDIA H200 NVL

from$4.63/hr

Available now · 2 regions

Memory: 141 GB HBM3e
FP16 tensor: 835 TFLOPS
Bandwidth: 4.8 TB/s

Launch H200 NVL Reserve a cluster

Full pricing details

CDNA 2

AMD MI250X

Not in stock right now

Memory: 128 GB HBM2e
FP16 tensor: 383 TFLOPS
Bandwidth: 3.3 TB/s

Full pricing details

At a glance

Spec by spec. The longer bar is the better number.

Memory

NVIDIA H200 NVL141 GB

AMD MI250X128 GB

Memory bandwidth

NVIDIA H200 NVL4.8 TB/s

AMD MI250X3.3 TB/s

FP16 tensor

NVIDIA H200 NVL835 TFLOPS

AMD MI250X383 TFLOPS

FP32

NVIDIA H200 NVL60 TFLOPS

AMD MI250X47.9 TFLOPS

Full specs

The complete spec sheet.

Spec	NVIDIA H200 NVL	AMD MI250X
Architecture	Hopper	CDNA 2
Memory	141 GB	128 GB
Memory type	HBM3e	HBM2e
Memory bandwidth	4.8 TB/s	3.3 TB/s
FP32	60 TFLOPS	47.9 TFLOPS
FP16 tensor	835 TFLOPS	383 TFLOPS
INT8	1670 TOPS	383 TOPS
TDP	600 W	560 W
Form factor	PCIe	OAM
NVLink	In groups of 4	No
GPUs per node	8	8
Hourly price (live)	from $4.63/hr	Not in stock right now

Questions

H200 NVL or MI250X. What buyers ask first.

Other comparisons