NVIDIA141 GB VRAM4,800 GB/s$4/hr

Qwen3.8 27B (VLM) on H200 SXM5 141GB: Does It Fit, How Fast, What It Costs

Qwen3.8 27B (VLM) at FP8 (E4M3) needs 27.8 GB for its weights and 33.3 GB in total at 8K context with 8 concurrent users. That means it fits on a single H200 SXM5 141GB. Every figure is computed by tokcalc's open-source capacity model.

Fits?
Yes, 1 GPU
33.3 GB needed / 141 GB available
Decode speed
168.5
tok/s per stream
Aggregate
2022
tok/s @ 8 concurrent
Cost / 1M tokens
$0.5496
1× H200 SXM5 141GB
What changes as you add H200 SXM5 141GBs
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× H200141 GB33.3 GB168.52022955.5 ms$0.5496✓
2× H200282 GB33.3 GB286.46873493.3 ms$0.3233✓
4× H200564 GB33.3 GB572.827493261.6 ms$0.1617✓
8× H2001128 GB33.3 GB1146109971145.8 ms$0.0808✓
Concurrency vs context length
On 1× H200 SXM5 141GB 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?
4K2691.07 GB✓ fits
8K1642.15 GB✓ fits
33K498.59 GB✓ fits
66K2517.18 GB✓ fits
131K1234.36 GB✓ fits
Other GPUs that run Qwen3.8 27B (VLM)
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
RTX 5090 32GB2×2566$0.1083
AMD MI300X 192GB1×2232$0.3111
A100 40GB SXM41×652.8$0.3830
AMD MI325X 288GB1×2527$0.3847
A100 80GB SXM41×842.3$0.3957
Intel Gaudi 3 128GB1×1390$0.3997
Frequently asked questions

Can Qwen3.8 27B (VLM) run on H200 SXM5 141GB?

Yes. Qwen3.8 27B (VLM) at FP8 (E4M3) needs 33.3 GB including the KV cache, and H200 SXM5 141GB has 141 GB — so it fits on a single GPU.

How fast is Qwen3.8 27B (VLM) on H200 SXM5 141GB?

168.5 tokens/sec per stream and 2022 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 955.5 ms, and inter-token latency is 5.9 ms.

What does Qwen3.8 27B (VLM) on H200 SXM5 141GB cost per million tokens?

$0.5496 per 1M output tokens at 1× H200 SXM5 141GB ($4.00/hr). That is the marginal hardware cost — add idle capacity, ops, and egress for the real figure.

How much VRAM does Qwen3.8 27B (VLM) need on H200 SXM5 141GB?

27.8 GB for the weights at FP8 (E4M3), plus 5.5 GB of KV cache for 8 concurrent requests at 8K context — 33.3 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is H200 SXM5 141GB the right choice for Qwen3.8 27B (VLM)?

It works. Compare $0.5496 per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 4× RTX 3080 10GB at $0.0511, and the fastest is 1× Cerebras CS-3 (WSE-3) at 8423 tok/s.

Adjust the workload yourself

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

Open in calculator