GPU model · price + stock · hourly snapshot
AMD MI300X cloud rental prices
1 provider lists the AMD MI300X, 1 has it in stock. It rents for $2.39 per GPU hour.
LIVE Prices and stock checked 1 minute ago
Providers
Every provider renting the AMD MI300X
| Provider | From | Stock | Grab |
|---|---|---|---|
| Runpod | $2.39 | In stock | Grab |
In stock first, then cheapest. Stock is what the provider reported at our last check. One GPU for a month (730h) at $2.39: $1,745.
The AMD MI300X sells out fast. Get one email when it is back in stock.
Price history
Price trend
$2.39 /GPU/hr · Oct 22
One point per day: the lowest listed $/GPU/hr across every provider we track, captured daily. 1 offers in the latest snapshot.
Configurations
The one configuration
| Configuration | VRAM | Per GPU/hr | Stock |
|---|---|---|---|
| 1x MI300XRunpod | 192 GB | $2.39 | In stock |
Model fit
What fits in 192 GB
The question behind most rentals: will the model load. Sized against one AMD MI300X.
| Model | 16-bit | 8-bit | 4-bit |
|---|---|---|---|
| DeepSeek-V3 DeepSeek · 671B params · 37B active | 1,610 GB 9× · 118 GB free | 805 GB 5× · 155 GB free | 403 GB 3× · 173 GB free |
| Llama 3.1 405B Meta · 405B params | 972 GB 6× · 180 GB free | 486 GB 3× · 90 GB free | 243 GB 2× · 141 GB free |
| Qwen3 235B-A22B Alibaba · 235B params · 22B active | 564 GB 3× · 12 GB free | 282 GB 2× · 102 GB free | 141 GB 1× · 51 GB free |
| Mixtral 8x22B Mistral · 141B params · 39B active | 338 GB 2× · 46 GB free | 169 GB 1× · 23 GB free | 85 GB 1× · 107 GB free |
| Mistral Large 2 Mistral · 123B params | 295 GB 2× · 89 GB free | 148 GB 1× · 44 GB free | 74 GB 1× · 118 GB free |
| gpt-oss-120b OpenAI · 117B params · 5.1B active | 281 GB 2× · 103 GB free | 140 GB 1× · 52 GB free | 70 GB 1× · 122 GB free |
| Qwen2.5 72B Alibaba · 72.7B params | 174 GB 1× · 18 GB free | 87 GB 1× · 105 GB free | 44 GB 1× · 148 GB free |
| Llama 3.3 70B Meta · 70.6B params | 169 GB 1× · 23 GB free | 85 GB 1× · 107 GB free | 42 GB 1× · 150 GB free |
| Qwen3 32B Alibaba · 32.8B params | 79 GB 1× · 113 GB free | 39 GB 1× · 153 GB free | 20 GB 1× · 172 GB free |
| Gemma 2 27B Google · 27.2B params | 65 GB 1× · 127 GB free | 33 GB 1× · 159 GB free | 16 GB 1× · 176 GB free |
| Mistral Small 3 Mistral · 24B params | 58 GB 1× · 134 GB free | 29 GB 1× · 163 GB free | 14 GB 1× · 178 GB free |
| Llama 3.1 8B Meta · 8B params | 19 GB 1× · 173 GB free | 10 GB 1× · 182 GB free | 5 GB 1× · 187 GB free |
Required memory is weights plus 20% for KV cache and runtime at short context; long contexts and large batches need more. Cells tinted green fit on a single AMD MI300X.
Specifications
AMD MI300X specifications
| Architecture | CDNA 3 |
|---|---|
| Memory per GPU | 192 GB HBM3 |
| Memory bandwidth | 5,300 GB/s |
| Release date | Q4 2023 |
| GPU interconnect | Infinity Fabric 896 GB/s |
| FP8 compute rate | 2,614.9 TFLOPS (dense), 5,229.8 with sparsity |
| FP16 / BF16 compute rate | 1,307.4 TFLOPS (dense), 2,614.9 with sparsity |
| INT8 compute rate | 2,614.9 TOPS (dense), 5,229.8 with sparsity |
| Process node | TSMC 5nm / 6nm FinFET |
| Board power | 750 W |
| Partitioning | Up to 8 partitions |
Alternatives