Can I deploy NVIDIA B300 or AMD MI250X on Runcrate today?

NVIDIA B300 is on demand in 1 region from $10.40/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 B300 has 288GB HBM3e. AMD MI250X has 128GB HBM2e.

B300 or MI250X: which holds a 70B model in 8-bit weights?

It needs 70 GB of 8-bit weights. On the B300: 1 GPU, 218 GB spare. On the MI250X: 1 GPU, 58 GB spare. Pick the B300: same GPU count, up to 2.4x the memory bandwidth.

GPU comparison

B300 vs MI250X.

The B300 is ahead on memory (288 GB vs 128 GB) and speed (5.9x the FP16 throughput and up to 2.4x the memory bandwidth). The MI250X makes sense when your model fits in 128 GB and its hourly price is lower.

Launch B300 Reserve a cluster

Rent either on Runcrate

B300from $10.40/hr
$0.036 per GB/hr$0.46 per 100 TFLOPS/hr
Launch B300
MI250XNot in stock right now
Register interest

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.

JobB300MI250XPick

Hold a 70B model in 8-bit weights
70 GB of 8-bit weights
B300: 1 GPU, 218 GB spare
MI250X: 1 GPU, 58 GB spare
B300. Same GPU count, up to 2.4x the memory bandwidth.
Serve an 8B model in 16-bit weights
16 GB of 16-bit weights
B300: 1 GPU, 272 GB spare
MI250X: 1 GPU, 112 GB spare
B300. Same GPU count, up to 2.4x the memory bandwidth.
Fine-tune a 70B model with QLoRA
About 46 GB with 4-bit weights
B300: 1 GPU, 242 GB spare
MI250X: 1 GPU, 82 GB spare
B300. Same GPU count, 5.9x the FP16 throughput.
Train on a full node
FP16 tensor throughput per node
B300: 8x, 18,000 TFLOPS, NVLink
MI250X: 8x, 3,064 TFLOPS
B300. 5.9x the FP16 throughput per GPU.
Throughput per watt
FP16 TFLOPS per 100 W of TDP
B300: Varies by card
MI250X: 68 TFLOPS per 100 W
Either. Varies by card.

Compare another pair

Blackwell

NVIDIA B300

from$10.40/hr

Available now · 1 region

Memory: 288 GB HBM3e
FP16 tensor: 2250 TFLOPS
Bandwidth: up to 8.0 TB/s

Launch B300 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 B300288 GB

AMD MI250X128 GB

Memory bandwidth

NVIDIA B300up to 8.0 TB/s

AMD MI250X3.3 TB/s

FP16 tensor

NVIDIA B3002250 TFLOPS

AMD MI250X383 TFLOPS

FP32

NVIDIA B30075 TFLOPS

AMD MI250X47.9 TFLOPS

Full specs

The complete spec sheet.

Spec	NVIDIA B300	AMD MI250X
Architecture	Blackwell	CDNA 2
Memory	288 GB	128 GB
Memory type	HBM3e	HBM2e
Memory bandwidth	up to 8.0 TB/s	3.3 TB/s
FP32	75 TFLOPS	47.9 TFLOPS
FP16 tensor	2250 TFLOPS	383 TFLOPS
INT8	187.5 TOPS	383 TOPS
TDP	up to 1400 W	560 W
Form factor	SXM	OAM
NVLink	Yes	No
GPUs per node	8	8
Hourly price (live)	from $10.40/hr	Not in stock right now

Questions

B300 or MI250X. What buyers ask first.

Other comparisons