On-demand, the MI355X rents for $8.60/hr and the MI300X for $3.10/hr — the MI300X is ~177% cheaper per hour, averaged across cloud providers.
| MI355X | MI300X | |
|---|---|---|
| On-demand /hr | $8.60 | $3.10 |
| Spot /hr | $— | $1.20 |
| Reserved /hr | $2.29 | $2.49 |
| Per month (on-demand) | $6,278 | $2,263 |
| MI355X | MI300X | |
|---|---|---|
| Vendor | AMD | AMD |
| Architecture | CDNA 4 | CDNA 3 |
| Memory | 288 GB HBM3e | 192 GB HBM3 |
| Memory bandwidth | 8 TB/s | 5.3 TB/s |
| Launched | 2025 | 2023 |
On-demand, the MI355X averages $8.60/hr vs the MI300X at $3.10/hr — the MI300X is about 177% cheaper per hour. Spot rates are $—/hr vs $1.20/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 MI300X is usually the better value.
At on-demand rates that's about $6,278/mo for the MI355X and $2,263/mo for the MI300X (730 hours).
See full pricing & history: MI355X price · MI300X price · all GPUs.
Cross-provider averages from public cloud GPU listings. Informational only.