GPU MODELS / MI325X
MI325X cloud GPU pricing
5 listings across 4 providers. Cheapest published rate: $3.800/GPU/hr at DigitalOcean. Sorted by per-GPU hourly price. Every row links to the provider's own published price.
WHAT IS THE MI325X
The MI325X is the memory-maximized refresh of the MI300X: same CDNA 3 compute, a third more HBM3e at 256 GB. It exists for one reason - holding larger models on one GPU - and where clouds list it, that is exactly the pitch.
WHAT IT IS GOOD AT
WHO SHOULD NOT RENT THE MI325X
| Provider | Config | VRAM | Location | $/hr | $/GPU/hr | Source |
|---|---|---|---|---|---|---|
| MI325X × 1 | 256 GB | Not published | $1.590 | $1.590STALE | receipt | |
| MI325X × 8 | 256 GB | Not published | $16.000 | $2.000STALE | receipt | |
| MI325X (bare metal) × 8 | 256 GB | Not published | $16.000 | $2.000STALE | receipt | |
| MI325X × 1 | 192 GB | India | $3.175 | $3.175STALE | receipt | |
| MI325X × 1 | 256 GB | Not published | $3.800 | $3.800 | receipt |
THE MI325X MARKET, BY THE NUMBERS
4 providers publish 5 MI325X configurations on the panel today. Published per-GPU hourly rates run from $3.800 (DigitalOcean) to $3.80 (DigitalOcean), a 1x spread between the cheapest and the most expensive published rate for the same chip. The median listing sits at $3.80/GPU/hr. Rows are published catalog rates; stock flags appear per row where the provider reports them. All rates are on-demand - committed-use and reserved pricing is excluded.
Across every MI325X row on the panel, the daily floor held flat at $3.800/GPU/hr over 52 days of tracking. Snapshots run daily; the full history is public in the repo.
At the current floor, one MI325X running around the clock costs about $2,774 a month (730 hours of on-demand arithmetic). Compare the alternatives below before committing - the same budget often buys more than one chip class.
MI325X VS THE ALTERNATIVES
Compare live rates: MI325X vs MI300X · MI325X vs H200 SXM · MI325X vs MI350X
MI325X PRICING FAQ
See also alternatives to the MI325X · cheapest GPU-hour per model · inference chips · training chips · by location