NVIDIA80 GB VRAM2,000 GB/s$1.2/hr

Gemma 2 27B on A100 80GB SXM4: 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 A100 80GB SXM4. Every figure is computed by tokcalc's open-source capacity model.

Fits?
Yes, 1 GPU
51.7 GB needed / 80 GB available
Decode speed
72.22
tok/s per stream
Aggregate
866.7
tok/s @ 8 concurrent
Cost / 1M tokens
$0.3846
1× A100 80GB SXM4
What changes as you add A100 80GB SXM4s
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× A10080 GB51.7 GB72.22866.72.88 s$0.3846✓
2× A100160 GB51.7 GB122.829471.46 s$0.2262✓
4× A100320 GB51.7 GB245.611787743.0 ms$0.1131✓
8× A100640 GB51.7 GB491.147147386.5 ms$0.0566✓
Concurrency vs context length
On 1× A100 80GB SXM4 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?
4K341.54 GB✓ fits
8K173.09 GB✓ fits
33K412.35 GB✓ fits
66K224.70 GB✓ fits
131K149.39 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
Intel Gaudi 3 128GB1×1430$0.3885
H100 SXM5 80GB1×1452$0.4784
Frequently asked questions

Can Gemma 2 27B run on A100 80GB SXM4?

Yes. Gemma 2 27B at FP8 (E4M3) needs 51.7 GB including the KV cache, and A100 80GB SXM4 has 80 GB — so it fits on a single GPU.

How fast is Gemma 2 27B on A100 80GB SXM4?

72.22 tokens/sec per stream and 866.7 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 2.88 s, and inter-token latency is 13.8 ms.

What does Gemma 2 27B on A100 80GB SXM4 cost per million tokens?

$0.3846 per 1M output tokens at 1× A100 80GB SXM4 ($1.20/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 A100 80GB SXM4?

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 A100 80GB SXM4 the right choice for Gemma 2 27B?

It works. Compare $0.3846 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