AMD192 GB VRAM5,300 GB/s$2.5/hr

Llama 3 8B on AMD MI300X 192GB: Does It Fit, How Fast, What It Costs

Llama 3 8B at FP16 needs 16.1 GB for its weights and 24.6 GB in total at 8K context with 8 concurrent users. That means it fits on a single AMD MI300X 192GB. Every figure is computed by tokcalc's open-source capacity model.

Fits?
Yes, 1 GPU
24.6 GB needed / 192 GB available
Decode speed
214.5
tok/s per stream
Aggregate
2574
tok/s @ 8 concurrent
Cost / 1M tokens
$0.2698
1× AMD MI300X 192GB
What changes as you add AMD MI300X 192GBs
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× AMD192 GB24.6 GB214.52574236.0 ms$0.2698✓
2× AMD384 GB24.6 GB364.78752133.4 ms$0.1587✓
4× AMD768 GB24.6 GB729.33500881.7 ms$0.0793✓
8× AMD1536 GB24.6 GB145914003155.9 ms$0.0397✓
Concurrency vs context length
On 1× AMD MI300X 192GB at FP16. KV cache scales linearly with context, so this is the number of users your GPUs can actually serve.
ContextConcurrent usersKV per requestFits?
4K327549.8 MB✓ fits
8K1631.07 GB✓ fits
33K404.29 GB✓ fits
66K208.59 GB✓ fits
131K1017.18 GB✓ fits
Other GPUs that run Llama 3 8B
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
RTX 5090 32GB1×870.3$0.1596
V100 SXM2 32GB1×437.1$0.3177
A100 40GB SXM41×752.8$0.3321
AMD MI325X 288GB1×2914$0.3336
A100 80GB SXM41×971.4$0.3432
Intel Gaudi 3 128GB1×1603$0.3466
Frequently asked questions

Can Llama 3 8B run on AMD MI300X 192GB?

Yes. Llama 3 8B at FP16 needs 24.6 GB including the KV cache, and AMD MI300X 192GB has 192 GB — so it fits on a single GPU.

How fast is Llama 3 8B on AMD MI300X 192GB?

214.5 tokens/sec per stream and 2574 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 236.0 ms, and inter-token latency is 4.7 ms.

What does Llama 3 8B on AMD MI300X 192GB cost per million tokens?

$0.2698 per 1M output tokens at 1× AMD MI300X 192GB ($2.50/hr). That is the marginal hardware cost — add idle capacity, ops, and egress for the real figure.

How much VRAM does Llama 3 8B need on AMD MI300X 192GB?

16.1 GB for the weights at FP16, plus 8.6 GB of KV cache for 8 concurrent requests at 8K context — 24.6 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is AMD MI300X 192GB the right choice for Llama 3 8B?

It works. Compare $0.2698 per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 4× RTX 3080 10GB at $0.0443, and the fastest is 1× Cerebras CS-3 (WSE-3) at 9714 tok/s.

Adjust the workload yourself

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

Open in calculator