On-demand, the MI355X rents for $8.60/hr and the MI325X for $3.26/hr — the MI325X is ~164% cheaper per hour, averaged across cloud providers.
| MI355X | MI325X | |
|---|---|---|
| On-demand /hr | $8.60 | $3.26 |
| Spot /hr | $— | $— |
| Reserved /hr | $2.29 | $1.98 |
| Per month (on-demand) | $6,278 | $2,380 |
| MI355X | MI325X | |
|---|---|---|
| Vendor | AMD | AMD |
| Architecture | CDNA 4 | CDNA 3 |
| Memory | 288 GB HBM3e | 256 GB HBM3e |
| Memory bandwidth | 8 TB/s | 6 TB/s |
| Launched | 2025 | 2025 |
On-demand, the MI355X averages $8.60/hr vs the MI325X at $3.26/hr — the MI325X is about 164% cheaper per hour. Spot rates are $—/hr vs $—/hr.
MI355X costs more per hour but offers 288 GB HBM3e memory and newer architecture — for large-model training the higher hourly rate often yields a lower cost-per-job. For memory-bound inference or budget runs, the MI325X is usually the better value.
At on-demand rates that's about $6,278/mo for the MI355X and $2,380/mo for the MI325X (730 hours).
See full pricing & history: MI355X price · MI325X price · all GPUs.
Cross-provider averages from public cloud GPU listings. Informational only.