NVIDIA32 GB VRAM1,792 GB/s$0.5/hr

Qwen 2.5-VL 72B (VLM) on RTX 5090 32GB: Does It Fit, How Fast, What It Costs

Qwen 2.5-VL 72B (VLM) at FP8 (E4M3) needs 72.0 GB for its weights and 93.5 GB in total at 8K context with 8 concurrent users. That means it needs 4× RTX 5090 32GB. Every figure is computed by tokcalc's open-source capacity model.

Fits?
Yes, 4×
93.5 GB needed / 128 GB available
Decode speed
82.51
tok/s per stream
Aggregate
3960
tok/s @ 8 concurrent
Cost / 1M tokens
$0.1403
4× RTX 5090 32GB

RTX 5090 32GB has no NVLink, so 4-way tensor parallelism runs over PCIe. tokcalc applies a flat 0.85× multi-GPU efficiency factor for this, which is realistic for NVLink-equipped datacenter cards but optimistic for PCIe scale-out — treat the throughput and cost figures for this configuration as an upper bound, and measure before committing.

What changes as you add RTX 5090 32GBs
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× RTX32 GB93.5 GB0.0000.000——✗
2× RTX64 GB93.5 GB0.0000.000——✗
4× RTX128 GB93.5 GB82.5139605.66 s$0.1403✓
8× RTX256 GB93.5 GB165.0158412.84 s$0.0701✓
Concurrency vs context length
On 4× RTX 5090 32GB 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?
4K411.34 GB✓ fits
8K202.68 GB✓ fits
33K510.74 GB✓ fits
66K221.47 GB✓ fits
131K142.95 GB✓ fits
Other GPUs that run Qwen 2.5-VL 72B (VLM)
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
A100 80GB SXM42×1105$0.6033
H100 SXM5 80GB2×1851$0.7504
AMD MI300X 192GB1×861.3$0.8063
L40S 48GB2×477.4$0.8147
AMD MI325X 288GB1×975.0$0.9972
Intel Gaudi 3 128GB1×536.3$1.04
Frequently asked questions

Can Qwen 2.5-VL 72B (VLM) run on RTX 5090 32GB?

Yes, but it needs 4× RTX 5090 32GB. At FP8 (E4M3) the total footprint is 93.5 GB (72.0 GB of weights plus 21.5 GB of KV cache at 8K × 8 concurrent users), against 128 GB of available VRAM across 4 GPUs.

How fast is Qwen 2.5-VL 72B (VLM) on RTX 5090 32GB?

82.51 tokens/sec per stream and 3960 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 5.66 s, and inter-token latency is 12.1 ms.

What does Qwen 2.5-VL 72B (VLM) on RTX 5090 32GB cost per million tokens?

$0.1403 per 1M output tokens at 4× RTX 5090 32GB ($0.5000/hr). That is the marginal hardware cost — add idle capacity, ops, and egress for the real figure.

How much VRAM does Qwen 2.5-VL 72B (VLM) need on RTX 5090 32GB?

72.0 GB for the weights at FP8 (E4M3), plus 21.5 GB of KV cache for 8 concurrent requests at 8K context — 93.5 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is RTX 5090 32GB the right choice for Qwen 2.5-VL 72B (VLM)?

It works. Compare $0.1403 per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 8× RTX 4080 16GB at $0.1052, and the fastest is 4× Cerebras CS-3 (WSE-3) at 44200 tok/s.

Adjust the workload yourself

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

Open in calculator