USA
4 configs
4x
AMD MI250
Previous-gen AMD data center GPU for HPC and scientific computing.
- Arch
- CDNA2
- Memory
- 128 GB HBM2e
- Bandwidth
- 3.20 TB/s
- Released
- Q4 2021
Price history
Weekly median price per GPU per hour
Aggregating historical prices...
At a glance Updated 1 day ago
- Cheapest in stock
- --
- Median price (current)
- --
- 90-day trend
- --
- Coverage
- 1 provider
- 4 price points
Pricing & availability
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.
No providers match .
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. Verify before provisioning. More on how we price.
What can the MI250 run?
One MI250 has 128 GB of VRAM. In practice, that's enough memory for roughly 191B parameters at 4-bit or 53B at 16-bit, assuming a 32K context. Below are some open-weight LLMs, with the estimated memory and GPUs each one needs.
Single node ≤ 8
Multi-node
| Model | Memory (INT4 / FP4) | MI250s needed | Cost /hr | Cost /mo |
|---|---|---|---|---|
Qwen3.8-27B
27B
|
19 GB
|
4
|
–
|
–
|
Gemma 4 31B
30.7B
|
24 GB
|
4
|
–
|
–
|
GPT-OSS-120B
117B · 5.1B active
|
68 GB
|
4
|
–
|
–
|
GLM-5.3-Flash
320B · 18B active
|
178 GB
|
4
|
–
|
–
|
MiniMax-M3
428B · 23B active
|
241 GB
|
4
|
–
|
–
|
DeepSeek V4.1 Flash
552B · 16B active
|
306 GB
|
4
|
–
|
–
|
Kimi K3
2.8T · 104B active
|
1,545 GB
|
14
|
–
|
–
|
Estimates based on the median on-demand rate. Memory is weights plus FP16/BF16 KV cache at 32K context per request. GPU counts assume 90% of advertised VRAM is usable. No guarantee of runtime support, usable performance, or that a matching quantized build exists. How we estimate costs.
Technical Specifications
AMD MI250 · Per GPU
| Compute · dense | |
|---|---|
| FP16 / BF16 | 362.1 TFLOPS |
| INT8 | 362.1 TOPS |
| FP32 | 90.5 TFLOPS |
| FP64 | 45.3 TFLOPS |
| Precision support | FP16BF16FP32FP64INT4INT8 |
| Memory | |
|---|---|
| Capacity | 128 GB HBM2e |
| Bandwidth | 3,200 GB/s |
| Bus width | 8,192-bit |
| ECC | Yes |
| Silicon | |
|---|---|
| Architecture | CDNA 2 |
| Process | TSMC 6nm FinFET |
| Shader cores | 13,312 Stream processors |
| Matrix cores | 832 |
| Compute units | 208 |
| Fabric and host | |
|---|---|
| GPU interconnect | Infinity Fabric 400 GB/s |
| Host interface | PCIe 4.0 x16 |
| Power | |
|---|---|
| Board power | 500 W |
| Cooling | Passive |
Source: official AMD MI250 datasheet.
Frequently Asked Questions
How much does the MI250 cost per hour?
-
As of September 12, 2026, we track 4 configs from 1 provider, none of them on demand. Prices are per GPU per hour.
Billing type Configs Cheapest Reserved
4
$1.30 (12 mo)
No median for reserved: we only show this when at least 3 providers list the GPU on that billing type.
Where can I rent an MI250?
-
GetDeploying currently tracks MI250 configs from 1 provider. See the full price comparison above for every provider and config.
How many MI250 GPUs do I need?
-
One MI250 runs models up to roughly 191B parameters at 4-bit quantization or 53B at 16-bit, assuming a 32K context. Larger models run across multiple GPUs: the model table above shows the estimated memory and GPU count for popular open-weight LLMs.
Why choose the MI250?
-
128GB HBM2e across dual GCDs. FP64 performance makes it strong for scientific HPC workloads. Competitive pricing as a previous-generation AMD data center GPU.
When is the MI250 not a good fit?
-
Older CDNA 2 architecture with no FP8 support. ROCm maturity for AI workloads, while improving, lags behind MI300X. Limited cloud provider availability.
Alternatives to AMD MI250
Last updated