AMD MI355X
🏢AMD's Blackwell-generation answer. 288 GB.
Prices collected 2026-09-12 · no listings in the latest snapshot
Compare vs other GPUs →At a glance
AMD MI355X pricing by provider
One card per provider, cheapest first. A configuration is one priced unit — an instance type, a SKU, a marketplace offer — and the per-GPU price is that unit's price divided by its GPU count. Open a card to see every configuration with its vCPU, RAM and region where the provider publishes them.
No provider we collect from listed this GPU on 2026-09-12.
What fits on a AMD MI355X
Weights + KV cache at 8K context + runtime overhead, against 259 GB usable of the 288 GB on the card. Capacity follows total parameters, even for mixture-of-experts models — any expert may be needed next. Where one card is not enough, the table says how many are, and prices the set at today's median.
| Model | Params | 4-bit | 8-bit | 16-bit |
|---|---|---|---|---|
| 2.8T104B active | 7× GPU1581 GB | 12× GPU2981 GB | 22× GPU5606 GB | |
| 552B16B active | 2× GPU313.5 GB | 3× GPU589.5 GB | 5× GPU1107 GB | |
 Llama 4 Maverick | 400B17B active | ✓ 228 GB~669 tok/s | 2× GPU428 GB | 4× GPU803 GB |
 gpt-oss-120b | 117B5.1B active | ✓ 67.9 GB~2231 tok/s | ✓ 126.4 GB~1181 tok/s | ✓ 236.1 GB~627 tok/s |
| Llama 3.1 70B | 70B | ✓ 44.5 GB~163 tok/s | ✓ 79.5 GB~86 tok/s | ✓ 145.1 GB~46 tok/s |
 Gemma 4 31B | 31B | ✓ 21.2 GB~367 tok/s | ✓ 36.7 GB~194 tok/s | ✓ 65.7 GB~103 tok/s |
 Qwen3 Coder 30B-A3B | 30.5B3.3B active | ✓ 19 GB~3448 tok/s | ✓ 34.3 GB~1825 tok/s | ✓ 62.9 GB~970 tok/s |
| Qwen3.8 27B | 27B | ✓ 18.7 GB~421 tok/s | ✓ 32.2 GB~223 tok/s | ✓ 57.6 GB~119 tok/s |
| gpt-oss-20b | 21B3.6B active | ✓ 13.7 GB~3160 tok/s | ✓ 24.2 GB~1673 tok/s | ✓ 43.9 GB~889 tok/s |
| Llama 3.1 8B | 8B | ✓ 6.9 GB~1422 tok/s | ✓ 10.9 GB~753 tok/s | ✓ 18.4 GB~400 tok/s |
tok/s is decode throughput from memory bandwidth, derated 20%, single stream — an estimate, not a benchmark. Set cost uses today's median × GPUs required.
Specifications
Bandwidth is listed because it governs generation speed: each token streams the active weights once, so tokens per second is bandwidth divided by bytes read. TFLOPs decide prompt processing and training, not how fast text appears.
Alternatives
Questions
How much does it cost to rent a AMD MI355X?
No provider we collect from listed the AMD MI355X on 2026-09-12. It stays in the catalogue and is priced the day one does.
Why do prices for the same GPU vary so much?
Hyperscalers bundle CPU, RAM and network into an instance price and charge for the ecosystem around it. Neoclouds sell the GPU more directly. Marketplaces are independent hosts competing on price, with the trade-offs of shared, third-party hardware. The per-GPU figure is the instance price divided by GPU count, which is the only way to compare an 8-GPU node with a single-card listing.
Is the median the price I will pay?
No — it is the middle of the market. Each provider's own median is taken first so a provider with many instance sizes counts once, then the median across providers. Spot and reserved rates are excluded from it; they set the floor shown separately. Use it to judge whether a quote is high or low, then verify with the provider.
Which models can a AMD MI355X run?
With 288 GB per GPU, the table above shows which reference models fit on one card at 4, 8 and 16-bit, using the same memory arithmetic as the rest of Runyard: weights plus KV cache plus runtime overhead, with capacity following total parameters even for mixture-of-experts models. Where a model needs more than one GPU it says how many and what that set costs at today's median.