NVIDIA48 GB VRAM864 GB/s$0.7/hr

Llama 3 8B on L40S 48GB: 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 L40S 48GB. Every figure is computed by tokcalc's open-source capacity model.

Fits?
Yes, 1 GPU
24.6 GB needed / 48 GB available
Decode speed
34.97
tok/s per stream
Aggregate
419.6
tok/s @ 8 concurrent
Cost / 1M tokens
$0.4634
1× L40S 48GB
What changes as you add L40S 48GBs
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× L40S48 GB24.6 GB34.97419.62.93 s$0.4634✓
2× L40S96 GB24.6 GB59.4514271.48 s$0.2726✓
4× L40S192 GB24.6 GB118.95707756.6 ms$0.1363✓
8× L40S384 GB24.6 GB237.822828393.3 ms$0.0681✓
Concurrency vs context length
On 1× L40S 48GB at FP16. KV cache scales linearly with context, so this is the number of users your GPUs can actually serve.
ContextConcurrent usersKV per requestFits?
4K59549.8 MB✓ fits
8K291.07 GB✓ fits
33K74.29 GB✓ fits
66K38.59 GB✓ fits
131K117.18 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
AMD MI300X 192GB1×2574$0.2698
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
Frequently asked questions

Can Llama 3 8B run on L40S 48GB?

Yes. Llama 3 8B at FP16 needs 24.6 GB including the KV cache, and L40S 48GB has 48 GB — so it fits on a single GPU.

How fast is Llama 3 8B on L40S 48GB?

34.97 tokens/sec per stream and 419.6 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 2.93 s, and inter-token latency is 28.6 ms.

What does Llama 3 8B on L40S 48GB cost per million tokens?

$0.4634 per 1M output tokens at 1× L40S 48GB ($0.7000/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 L40S 48GB?

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 L40S 48GB the right choice for Llama 3 8B?

It works. Compare $0.4634 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