AMD MI250 vs Nvidia T4

The MI250 is newer, with 8x the memory and 11x the bandwidth. 1 provider lists it, against 10 for the T4.

At a glance

Memory

8x more on MI250

MI250: 128 GB

T4: 16 GB

Memory bandwidth

11x higher on MI250

MI250: 3,200 GB/s

T4: 300 GB/s

Cheapest on-demand

MI250 is not sold on demand

MI250: --

T4: $0.20 /GPU/hr

Price comparison

Median price per GPU per hour by billing type, with the cheapest listing under it.

Billing	MI250	T4
On-demand	No on-demand listings	$0.69 /GPU/hr from $0.20
Reserved	-- /GPU/hr from $1.30 (12 mo)	$0.43 /GPU/hr from $0.20 (36 mo)
Spot	No spot listings	-- /GPU/hr from $0.15

Heads up: A provider's own page may quote a different figure, for example on tax, region or a promotion. A monthly-only plan shows a derived hourly rate. Medians are across providers, one vote each, so a large catalog does not outweigh a small one. No median for reserved on the MI250 and spot on the T4: we only show one when at least 3 providers list the GPU on that billing type. Verify before provisioning. More on how we price.

Price history

MI250 vs T4 price history

Weekly median on-demand price per GPU per hour

Last 12 months

Aggregating historical prices...

Specs comparison

Spec	MI250	T4	Difference
Memory
Capacity	128 GB HBM2e	16 GB GDDR6	8x more on MI250
Bandwidth	3,200 GB/s	300 GB/s	11x higher on MI250
Bus width	8,192-bit	256-bit	32x wider on MI250
Compute · dense
FP16 / BF16	362.1 TFLOPS	65 TFLOPS	5.6x higher on MI250
INT8	362.1 TOPS	130 TOPS	2.8x higher on MI250
FP32	90.5 TFLOPS	8.1 TFLOPS	11x higher on MI250
FP64	45.3 TFLOPS	Not supported	MI250 only
Platform
Interconnect	Infinity Fabric 400 GB/s	PCIe only	Infinity Fabric on MI250 only
Board power	500 W	70 W	7.1x higher on MI250
Released	Q4 2021	Q3 2018	3 years newer on MI250
Precision
Supported formats	FP16BF16FP32FP64INT4INT8	FP16FP32INT4INT8

Compute: Vendor peak figures, dense and per GPU. Not measured throughput.

Source: MI250 datasheet and T4 datasheet.

Where to rent them

Providers that rent these GPUs, with each one's cheapest rate. No provider we track offers both right now.

Provider	MI250	T4
Cirrascale USA	$1.30 /GPU/hr Reserved	--
Zenlayer USA	--	$0.56 /GPU/hr On-demand PCIe
Theta EdgeCloud Distributed	--	$0.20 /GPU/hr On-demand
Hugging Face USA	--	$0.40 /GPU/hr On-demand
Google Cloud USA	--	$0.42 /GPU/hr On-demand
AWS USA	--	$0.53 /GPU/hr On-demand
Azure USA	--	$0.53 /GPU/hr On-demand
Cerebrium USA	--	$0.59 /GPU/hr On-demand
Replicate USA	--	$0.81 /GPU/hr On-demand