Google95 GB VRAM819 GB/s$4.2/hr

Mixtral 8x7B (MoE) on TPU v5p (per-chip, 95GB): Does It Fit, How Fast, What It Costs

Mixtral 8x7B (MoE) at FP8 (E4M3) needs 46.7 GB for its weights and 55.3 GB in total at 8K context with 8 concurrent users. That means it fits on a single TPU v5p (per-chip, 95GB). Every figure is computed by tokcalc's open-source capacity model.

Fits?
Yes, 1 GPU
55.3 GB needed / 95 GB available
Decode speed
61.90
tok/s per stream
Aggregate
742.8
tok/s @ 8 concurrent
Cost / 1M tokens
$1.57
1× TPU v5p (per-chip, 95GB)
What changes as you add TPU v5p (per-chip, 95GB)s
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× TPU95 GB55.3 GB61.90742.8967.1 ms$1.57✓
2× TPU190 GB55.3 GB105.22526500.0 ms$0.9239✓
4× TPU380 GB55.3 GB210.510102265.0 ms$0.4619✓
8× TPU760 GB55.3 GB420.940409147.5 ms$0.2310✓
Concurrency vs context length
On 1× TPU v5p (per-chip, 95GB) 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?
4K89549.8 MB✓ fits
8K441.07 GB✓ fits
33K114.29 GB✓ fits
66K58.59 GB✓ fits
131K217.18 GB✓ fits
Other GPUs that run Mixtral 8x7B (MoE)
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
RTX 5090 32GB2×5526$0.0503
AMD MI300X 192GB1×4807$0.1445
L40S 48GB2×2664$0.1460
AMD MI325X 288GB1×5442$0.1787
A100 80GB SXM41×1814$0.1838
Intel Gaudi 3 128GB1×2993$0.1856
Frequently asked questions

Can Mixtral 8x7B (MoE) run on TPU v5p (per-chip, 95GB)?

Yes. Mixtral 8x7B (MoE) at FP8 (E4M3) needs 55.3 GB including the KV cache, and TPU v5p (per-chip, 95GB) has 95 GB — so it fits on a single GPU.

How fast is Mixtral 8x7B (MoE) on TPU v5p (per-chip, 95GB)?

61.90 tokens/sec per stream and 742.8 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 967.1 ms, and inter-token latency is 16.2 ms.

What does Mixtral 8x7B (MoE) on TPU v5p (per-chip, 95GB) cost per million tokens?

$1.57 per 1M output tokens at 1× TPU v5p (per-chip, 95GB) ($4.20/hr). That is the marginal hardware cost — add idle capacity, ops, and egress for the real figure.

How much VRAM does Mixtral 8x7B (MoE) need on TPU v5p (per-chip, 95GB)?

46.7 GB for the weights at FP8 (E4M3), plus 8.6 GB of KV cache for 8 concurrent requests at 8K context — 55.3 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is TPU v5p (per-chip, 95GB) the right choice for Mixtral 8x7B (MoE)?

It works. Compare $1.57 per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 8× RTX 3080 10GB at $0.0119, and the fastest is 2× Cerebras CS-3 (WSE-3) at 61674 tok/s.

Adjust the workload yourself

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

Open in calculator