NVIDIA80 GB VRAM2,000 GB/s$1.2/hr

Llama 4 Maverick 17B-128E (MoE, VLM) on A100 80GB SXM4: Does It Fit, How Fast, What It Costs

Llama 4 Maverick 17B-128E (MoE, VLM) at AWQ 4-bit needs 220.0 GB for its weights and 232.9 GB in total at 8K context with 8 concurrent users. That means it needs 4× A100 80GB SXM4. Every figure is computed by tokcalc's open-source capacity model.

Fits?
Yes, 4×
232.9 GB needed / 320 GB available
Decode speed
416.0
tok/s per stream
Aggregate
19968
tok/s @ 8 concurrent
Cost / 1M tokens
$0.0668
4× A100 80GB SXM4
What changes as you add A100 80GB SXM4s
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× A10080 GB232.9 GB0.0000.000——✗
2× A100160 GB232.9 GB0.0000.000——✗
4× A100320 GB232.9 GB416.019968478.8 ms$0.0668✓
8× A100640 GB232.9 GB832.079872254.4 ms$0.0334✓
Concurrency vs context length
On 4× A100 80GB SXM4 at AWQ 4-bit. KV cache scales linearly with context, so this is the number of users your GPUs can actually serve.
ContextConcurrent usersKV per requestFits?
4K124824.6 MB✓ fits
8K621.61 GB✓ fits
33K156.44 GB✓ fits
66K712.88 GB✓ fits
131K325.77 GB✓ fits
Other GPUs that run Llama 4 Maverick 17B-128E (MoE, VLM)
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
RTX 5090 32GB8×71565$0.0155
L40S 48GB8×34505$0.0451
H100 SXM5 80GB4×33446$0.0831
B200 SXM 192GB2×19968$0.1669
H200 SXM5 141GB2×11981$0.1855
AMD MI325X 288GB1×4405$0.2207
Frequently asked questions

Can Llama 4 Maverick 17B-128E (MoE, VLM) run on A100 80GB SXM4?

Yes, but it needs 4× A100 80GB SXM4. At AWQ 4-bit the total footprint is 232.9 GB (220.0 GB of weights plus 12.9 GB of KV cache at 8K × 8 concurrent users), against 320 GB of available VRAM across 4 GPUs.

How fast is Llama 4 Maverick 17B-128E (MoE, VLM) on A100 80GB SXM4?

416.0 tokens/sec per stream and 19968 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 478.8 ms, and inter-token latency is 2.4 ms.

What does Llama 4 Maverick 17B-128E (MoE, VLM) on A100 80GB SXM4 cost per million tokens?

$0.0668 per 1M output tokens at 4× A100 80GB SXM4 ($1.20/hr). That is the marginal hardware cost — add idle capacity, ops, and egress for the real figure.

How much VRAM does Llama 4 Maverick 17B-128E (MoE, VLM) need on A100 80GB SXM4?

220.0 GB for the weights at AWQ 4-bit, plus 12.9 GB of KV cache for 8 concurrent requests at 8K context — 232.9 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is A100 80GB SXM4 the right choice for Llama 4 Maverick 17B-128E (MoE, VLM)?

It works. Compare $0.0668 per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 8× RTX 5090 32GB at $0.0155, and the fastest is 8× Cerebras CS-3 (WSE-3) at 798720 tok/s.

Adjust the workload yourself

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

Open in calculator