AMD288 GB VRAM6,000 GB/s$3.5/hr

Mixtral 8x7B (MoE) on AMD MI325X 288GB: Does It Fit, How Fast, What It Costs

Mixtral 8x7B (MoE) at FP8 (E4M3) needs 46.7 GB for its weights and 55.3 GB in total at 8K context with 8 concurrent users. That means it fits on a single AMD MI325X 288GB. Every figure is computed by tokcalc's open-source capacity model.

Fits?
Yes, 1 GPU
55.3 GB needed / 288 GB available
Decode speed
453.5
tok/s per stream
Aggregate
5442
tok/s @ 8 concurrent
Cost / 1M tokens
$0.1787
1× AMD MI325X 288GB
What changes as you add AMD MI325X 288GBs
Tensor-parallel degree vs VRAM, throughput, and cost. Cost scales linearly with GPU count; decode throughput scales near-linearly too, so cost per token stays roughly flat.
GPUsTotal VRAMVRAM neededDecode tok/sAggregate tok/sTTFTCost / 1M tokFits?
1× AMD288 GB55.3 GB453.55442355.6 ms$0.1787✓
2× AMD576 GB55.3 GB770.918502193.0 ms$0.1051✓
4× AMD1152 GB55.3 GB154274009111.5 ms$0.0525✓
8× AMD2304 GB55.3 GB308429603770.8 ms$0.0263✓
Concurrency vs context length
On 1× AMD MI325X 288GB at FP8 (E4M3). KV cache scales linearly with context, so this is the number of users your GPUs can actually serve.
ContextConcurrent usersKV per requestFits?
4K449549.8 MB✓ fits
8K2241.07 GB✓ fits
33K564.29 GB✓ fits
66K288.59 GB✓ fits
131K1417.18 GB✓ fits
Other GPUs that run Mixtral 8x7B (MoE)
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
RTX 5090 32GB2×5526$0.0503
AMD MI300X 192GB1×4807$0.1445
L40S 48GB2×2664$0.1460
A100 80GB SXM41×1814$0.1838
Intel Gaudi 3 128GB1×2993$0.1856
H100 SXM5 80GB1×3038$0.2286
Frequently asked questions

Can Mixtral 8x7B (MoE) run on AMD MI325X 288GB?

Yes. Mixtral 8x7B (MoE) at FP8 (E4M3) needs 55.3 GB including the KV cache, and AMD MI325X 288GB has 288 GB — so it fits on a single GPU.

How fast is Mixtral 8x7B (MoE) on AMD MI325X 288GB?

453.5 tokens/sec per stream and 5442 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 355.6 ms, and inter-token latency is 2.2 ms.

What does Mixtral 8x7B (MoE) on AMD MI325X 288GB cost per million tokens?

$0.1787 per 1M output tokens at 1× AMD MI325X 288GB ($3.50/hr). That is the marginal hardware cost — add idle capacity, ops, and egress for the real figure.

How much VRAM does Mixtral 8x7B (MoE) need on AMD MI325X 288GB?

46.7 GB for the weights at FP8 (E4M3), plus 8.6 GB of KV cache for 8 concurrent requests at 8K context — 55.3 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is AMD MI325X 288GB the right choice for Mixtral 8x7B (MoE)?

It works. Compare $0.1787 per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 8× RTX 3080 10GB at $0.0119, and the fastest is 2× Cerebras CS-3 (WSE-3) at 61674 tok/s.

Adjust the workload yourself

Same formulas, your context length, concurrency, and quantisation.

Open in calculator