NVIDIA48 GB VRAM864 GB/s$0.7/hr

Qwen 3 4B on L40S 48GB: Does It Fit, How Fast, What It Costs

Qwen 3 4B at FP16 needs 8.0 GB for its weights and 17.7 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
17.7 GB needed / 48 GB available
Decode speed
70.20
tok/s per stream
Aggregate
842.4
tok/s @ 8 concurrent
Cost / 1M tokens
$0.2308
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 GB17.7 GB70.20842.41.48 s$0.2308✓
2× L40S96 GB17.7 GB119.32864753.8 ms$0.1358✓
4× L40S192 GB17.7 GB238.711457391.9 ms$0.0679✓
8× L40S384 GB17.7 GB477.445827211.0 ms$0.0339✓
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?
4K66618.5 MB✓ fits
8K331.21 GB✓ fits
33K84.83 GB✓ fits
66K49.66 GB✓ fits
131K219.33 GB✓ fits
Other GPUs that run Qwen 3 4B
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
RTX 3090 24GB1×912.6$0.0761
RTX 5090 32GB1×1747$0.0795
RTX 4090 24GB1×982.8$0.1131
AMD MI300X 192GB1×5168$0.1344
V100 SXM2 32GB1×877.5$0.1583
A100 40GB SXM41×1511$0.1654
Frequently asked questions

Can Qwen 3 4B run on L40S 48GB?

Yes. Qwen 3 4B at FP16 needs 17.7 GB including the KV cache, and L40S 48GB has 48 GB — so it fits on a single GPU.

How fast is Qwen 3 4B on L40S 48GB?

70.20 tokens/sec per stream and 842.4 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 1.48 s, and inter-token latency is 14.2 ms.

What does Qwen 3 4B on L40S 48GB cost per million tokens?

$0.2308 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 Qwen 3 4B need on L40S 48GB?

8.0 GB for the weights at FP16, plus 9.7 GB of KV cache for 8 concurrent requests at 8K context — 17.7 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is L40S 48GB the right choice for Qwen 3 4B?

It works. Compare $0.2308 per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 2× RTX 3080 10GB at $0.0441, and the fastest is 1× Cerebras CS-3 (WSE-3) at 19500 tok/s.

Adjust the workload yourself

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

Open in calculator