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

Gemma 2 27B on H200 SXM5 141GB: Does It Fit, How Fast, What It Costs

Gemma 2 27B at FP8 (E4M3) needs 27.0 GB for its weights and 51.7 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
51.7 GB needed / 141 GB available
Decode speed
173.3
tok/s per stream
Aggregate
2080
tok/s @ 8 concurrent
Cost / 1M tokens
$0.5342
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 GB51.7 GB173.32080929.4 ms$0.5342✓
2× H200282 GB51.7 GB294.77072480.2 ms$0.3142✓
4× H200564 GB51.7 GB589.328288255.1 ms$0.1571✓
8× H2001128 GB51.7 GB1179113152142.6 ms$0.0786✓
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?
4K731.54 GB✓ fits
8K363.09 GB✓ fits
33K912.35 GB✓ fits
66K424.70 GB✓ fits
131K249.39 GB✓ fits
Other GPUs that run Gemma 2 27B
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
RTX 5090 32GB2×2640$0.1052
AMD MI300X 192GB1×2297$0.3024
L40S 48GB2×1273$0.3055
AMD MI325X 288GB1×2600$0.3739
A100 80GB SXM41×866.7$0.3846
Intel Gaudi 3 128GB1×1430$0.3885
Frequently asked questions

Can Gemma 2 27B run on H200 SXM5 141GB?

Yes. Gemma 2 27B at FP8 (E4M3) needs 51.7 GB including the KV cache, and H200 SXM5 141GB has 141 GB — so it fits on a single GPU.

How fast is Gemma 2 27B on H200 SXM5 141GB?

173.3 tokens/sec per stream and 2080 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 929.4 ms, and inter-token latency is 5.8 ms.

What does Gemma 2 27B on H200 SXM5 141GB cost per million tokens?

$0.5342 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 Gemma 2 27B need on H200 SXM5 141GB?

27.0 GB for the weights at FP8 (E4M3), plus 24.7 GB of KV cache for 8 concurrent requests at 8K context — 51.7 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is H200 SXM5 141GB the right choice for Gemma 2 27B?

It works. Compare $0.5342 per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 8× RTX 3080 10GB at $0.0248, and the fastest is 2× Cerebras CS-3 (WSE-3) at 29467 tok/s.

Adjust the workload yourself

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

Open in calculator