Compare XPUs

Select up to 5 XPUs to compare side-by-side

Select XPUs to Compare

Clear all (5)Jump to results

Filter by Vendor

Showing 72 XPUs • 5 selected

Alibaba

Hanguang 800

AMD

MI100

23.1 TFLOPs

AMD

MI210

181 TFLOPs

AMD

MI250X

383 TFLOPs

AMD

MI300A

980.6 TFLOPs

AMD

MI300X

1,307 TFLOPs

AMD

MI325X

1,400 TFLOPs

AMD

MI350X

2,100 TFLOPs

AMD

Radeon PRO W7900

122 TFLOPs

AMD

Radeon RX 7900 XT

104 TFLOPs

AMD

Radeon RX 7900 XTX

122 TFLOPs

AWS

Inferentia2

190 TFLOPs

AWS

Trainium

190 TFLOPs

AWS

Trainium2

680 TFLOPs

Baidu

Kunlun II

Biren Technology

BR100

Cambricon

MLU370

256 TFLOPs

Cerebras

WSE-3

Enflame Technology

CloudBlazer T20

Etched

Sohu

10,000 TFLOPs

FuriosaAI

Warboy

Google

TPU v4

275 TFLOPs

Google

TPU v5e

197 TFLOPs

Google

TPU v5p

459 TFLOPs

Google

TPU v6e (Trillium)

918 TFLOPs

Graphcore

Bow IPU

Graphcore

IPU-M2000

Groq

LPU Inference Engine

Huawei

Ascend 910B

Iluvatar CoreX

BI-V150

300 TFLOPs

Intel

Data Center GPU Max 1100

177 TFLOPs

Intel

Data Center GPU Max 1550

419 TFLOPs

Intel Habana

Gaudi 2

432 TFLOPs

Intel Habana

Gaudi 3

1,835 TFLOPs

Multi-Metric Comparison

Relative performance across 5 key metrics (normalized to 100 = best in comparison)

Compute Performance (BF16)

Memory Capacity

Power Consumption

Power Efficiency

Specifications

Specification	NVIDIA B200	AMD MI350X	Cerebras WSE-3	Meta MTIA v1	AWS Trainium
Architecture	Blackwell	CDNA 4	Wafer Scale Engine	Meta Training & Inference Accelerator	Inferentia/Trainium
Form Factor	SXM	OAM	—	—	—
VRAM	192 GB	288 GB	44 GB	128 GB	32 GB
Memory Bandwidth	8,000 GB/s	—	—	—	—
TFLOPs (FP32)	90	—	—	—	—
TFLOPs (FP16)	2,250	—	—	—	—
TFLOPs	2,250	2,100	—	—	190
TFLOPs (FP8)	4,500	—	—	—	—
TDP	1000 W	850 W	23000 W	400 W	200 W
Launch Date	Mar 2024	Jun 2025	Mar 2024	May 2023	Nov 2021

Efficiency Metrics

Metric	B200	MI350X	WSE-3	MTIA v1	Trainium
TFLOPs per Watt (FP32-eq)	1.13	1.24	—	—	0.47
Memory Bandwidth per GB	41.7 GB/s	—	—	—	—

Performance Equivalence

How many units of each GPU are needed to match the performance of the others?

To match 1x NVIDIA B200

AMD MI350X

Compute (FP32-eq)

1.07x

Need 1.07x MI350X

VRAM

0.67x

MI350X has 1.50x more

Cerebras WSE-3

VRAM

4.36x

Need 4.36x WSE-3

Meta MTIA v1

VRAM

1.50x

Need 1.50x MTIA v1

AWS Trainium

Compute (FP32-eq)

11.84x

Need 11.84x Trainium

VRAM

6.00x

Need 6.00x Trainium

To match 1x AMD MI350X

NVIDIA B200

Compute (FP32-eq)

0.93x

B200 is 1.07x faster

VRAM

1.50x

Need 1.50x B200

Cerebras WSE-3

VRAM

6.55x

Need 6.55x WSE-3

Meta MTIA v1

VRAM

2.25x

Need 2.25x MTIA v1

AWS Trainium

Compute (FP32-eq)

11.05x

Need 11.05x Trainium

VRAM

9.00x

Need 9.00x Trainium

To match 1x Cerebras WSE-3

NVIDIA B200

VRAM

0.23x

B200 has 4.36x more

AMD MI350X

VRAM

0.15x

MI350X has 6.55x more

Meta MTIA v1

VRAM

0.34x

MTIA v1 has 2.91x more

AWS Trainium

VRAM

1.38x

Need 1.38x Trainium

To match 1x Meta MTIA v1

NVIDIA B200

VRAM

0.67x

B200 has 1.50x more

AMD MI350X

VRAM

0.44x

MI350X has 2.25x more

Cerebras WSE-3

VRAM

2.91x

Need 2.91x WSE-3

AWS Trainium

VRAM

4.00x

Need 4.00x Trainium

To match 1x AWS Trainium

NVIDIA B200

Compute (FP32-eq)

0.08x

B200 is 11.84x faster

VRAM

0.17x

B200 has 6.00x more

AMD MI350X

Compute (FP32-eq)

0.09x

MI350X is 11.05x faster

VRAM

0.11x

MI350X has 9.00x more

Cerebras WSE-3

VRAM

0.73x

WSE-3 has 1.38x more

Meta MTIA v1

VRAM

0.25x

MTIA v1 has 4.00x more

Pricing

Price Type	B200	MI350X	WSE-3	MTIA v1	Trainium
CAPEX (Street Price)	$70,000	—	—	—	—
OPEX (per hour)	—	—	—	—	—
Price per TFLOPs (FP32-eq)	$62	—	—	—	—