Why choose the MI210?

64GB HBM2e on a single 300W PCIe card, with FP64 matrix throughput that matches its FP32, which suits scientific HPC and simulation more than most accelerators at this price. Fits standard dual-slot servers rather than needing an OAM baseboard.

When is the MI210 not a good fit?

CDNA 2 has no FP8 support and half the memory of the dual-die MI250 , so large-model inference is tight. The ROCm software ecosystem, while rapidly improving, is narrower than CUDA. For AI-first work at a similar footprint, L40S has broader framework support.

What size AI models can the MI210 run?

With 64GB of VRAM, the MI210 can typically run models up to about 30B parameters in FP16, or 70B-class models in 4-bit quantized form for inference.

How much VRAM does the MI210 have?

The MI210 has 64GB of VRAM. Multi-GPU setups increase total memory, but that memory is not automatically pooled across GPUs.

What is the MI210's memory bandwidth?

The MI210 has 1,638 GB/s of memory bandwidth. Higher bandwidth helps with faster data transfer between GPU memory and compute cores.

What data types does the MI210 support?

The MI210 supports 6 precision formats. Training: BF16, FP16, FP32. Inference: INT4, INT8. Scientific: FP64.

Does the MI210 support Infinity Fabric?

Yes. The MI210 supports Infinity Fabric with 300 GB/s of bidirectional bandwidth. This helps accelerate multi-GPU communication.

How much does the MI210 cost per hour?

MI210 pricing currently ranges from $0.87/hr to $1.28/hr per GPU, depending on the provider, instance type, and billing model.

How much does the MI210 cost per month?

At 720 hours per month, one MI210 can cost between $626.40 to $921.60 per month, depending on the provider. Reserved and spot pricing can lower that further.

Which cloud providers offer the MI210?

The MI210 is available from 1 cloud provider: Leaseweb.

Can I rent the MI210 in the cloud?

Yes. We currently track 4 MI210 listings across 1 cloud providers: Billing type Listings Avg $/GPU/hr Reserved 4 $1.05/hr

AMD MI210

Single-die CDNA 2 data center GPU for HPC and double-precision workloads.

Latest update 5 hours ago

Compare vs other GPUs →

Aggregating historical prices...

Key Specifications

Architecture

CDNA2

Memory

64GB HBM2e

Memory Bandwidth

1,638 GB/s

Release date

Q1 2022

MI210 Pricing and Availability

Prices for the MI210 currently range from $0.87 to $1.28/hr per GPU. Depending on the provider, you can secure capacity at the lower end ($0.87/hr) to reduce hourly costs.

1 providers

How this list works

The two views

By provider shows one card per company, placed where its best offer ranked. By configuration lists every offer.

Order

What you can rent comes first: in stock, then waitlist, then not reported, then out of stock. Priced offers rank ahead of quote-only, and on-demand ahead of other billing types.

Within a group, five factors set the order:

Location: datacenter proximity, blended with provider HQ.
Price: hourly price, per GPU and in total.
Billing type: reserved ahead of spot, spot ahead of quote-only.
Specs: more VRAM, vCPUs and RAM.
Provider diversity: a provider's repeat rows rank slightly lower, so one company can't take all the top spots.

Search

We match the provider, its country, GPU model, form factor, billing type, availability and instance name. Matching is partial and case-insensitive. Everyday words work too, so "interruptible" finds spot and "sold out" finds out of stock.

Transparency and funding

Ads and sponsors: Paid placements sit at the top of the list and are always labeled as sponsored content. Sponsorship never influences the organic ranking itself.
Affiliates: Affiliate links are marked. We may earn a commission if you click them, but commissions never affect the order.
Prices: Shown in USD, converted at daily reference rates where a provider publishes in another currency. A month means 720 hours. How we estimate costs.

Leaseweb

Netherlands 4 configs 1x

Reserved from $0.87

From $0.87 / GPU / hr Reserved (1mo) Visit website


1x MI210 1x MI210 64GB (VOLUME_DL380G10-12LFF-5218-MI210)	64GB	32	128GB	Reserved (1mo)	$0.87	$0.87	Not reported We couldn't confirm availability. Please check with provider.
1x MI210 1x MI210 64GB (PREMIUM_R7525-16SFF8NVME-7413-MI210)	64GB	48	128GB	Reserved (1mo)	$0.97	$0.97	Not reported We couldn't confirm availability. Please check with provider.
1x MI210 1x MI210 64GB (PREMIUM_R7525-16SFF8NVME-7543-MI210)	64GB	64	128GB	Reserved (1mo)	$1.10	$1.10	Not reported We couldn't confirm availability. Please check with provider.
1x MI210 1x MI210 64GB (BUSINESS_DL380G11-8SFF-5520-MI210)	64GB	56	128GB	Reserved (1mo)	$1.28	$1.28	Not reported We couldn't confirm availability. Please check with provider.

Heads up: We do our best to keep these specs & prices accurate. However, cloud costs may fluctuate based on region, usage, and other factors not listed here. These are estimates based on common setups and are for informational purposes only. Always verify current rates & exact specs with the provider before provisioning.

Frequently Asked Questions

Why choose the MI210?: 64GB HBM2e on a single 300W PCIe card, with FP64 matrix throughput that matches its FP32, which suits scientific HPC and simulation more than most accelerators at this price. Fits standard dual-slot servers rather than needing an OAM baseboard.
When is the MI210 not a good fit?: CDNA 2 has no FP8 support and half the memory of the dual-die MI250, so large-model inference is tight. The ROCm software ecosystem, while rapidly improving, is narrower than CUDA. For AI-first work at a similar footprint, L40S has broader framework support.
What size AI models can the MI210 run?: With 64GB of VRAM, the MI210 can typically run models up to about 30B parameters in FP16, or 70B-class models in 4-bit quantized form for inference.
How much VRAM does the MI210 have?: The MI210 has 64GB of VRAM. Multi-GPU setups increase total memory, but that memory is not automatically pooled across GPUs.
What is the MI210's memory bandwidth?: The MI210 has 1,638 GB/s of memory bandwidth. Higher bandwidth helps with faster data transfer between GPU memory and compute cores.
What data types does the MI210 support?: The MI210 supports 6 precision formats. Training: BF16, FP16, FP32. Inference: INT4, INT8. Scientific: FP64.
Does the MI210 support Infinity Fabric?: Yes. The MI210 supports Infinity Fabric with 300 GB/s of bidirectional bandwidth. This helps accelerate multi-GPU communication.
How much does the MI210 cost per hour?: MI210 pricing currently ranges from $0.87/hr to $1.28/hr per GPU, depending on the provider, instance type, and billing model.
How much does the MI210 cost per month?: At 720 hours per month, one MI210 can cost between $626.40 to $921.60 per month, depending on the provider. Reserved and spot pricing can lower that further.
Which cloud providers offer the MI210?: The MI210 is available from 1 cloud provider: Leaseweb.
Can I rent the MI210 in the cloud?: Yes. We currently track 4 MI210 listings across 1 cloud providers:

Billing type Listings Avg $/GPU/hr

Reserved 4 $1.05/hr

Billing type	Listings	Avg $/GPU/hr
Reserved	4	$1.05/hr

Technical Specifications


GPU Architecture	CDNA2
Lithography	TSMC 6nm FinFET
Stream Processors	6,656
Compute Units	104
Peak FP16 Performance	181.0 TFLOPs
Peak BF16 Performance	181.0 TFLOPs
Peak FP32 Matrix Performance	45.3 TFLOPs
Peak FP64 Matrix Performance	45.3 TFLOPs
Peak FP32 Performance	22.6 TFLOPs
Peak FP64 Performance	22.6 TFLOPs
Peak INT8 Performance	181.0 TOPs
Peak INT4 Performance	181.0 TOPs
Form Factor	PCIe Gen4 x16, dual-slot
Board Power	300W