Google95 GB VRAM819 GB/s$4.2/hr

Qwen 2.5-VL 7B (VLM) on TPU v5p (per-chip, 95GB): Does It Fit, How Fast, What It Costs

Qwen 2.5-VL 7B (VLM) at FP16 needs 14.0 GB for its weights and 17.8 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
17.8 GB needed / 95 GB available
Decode speed
38.02
tok/s per stream
Aggregate
456.3
tok/s @ 8 concurrent
Cost / 1M tokens
$2.56
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 GB17.8 GB38.02456.3556.0 ms$2.56✓
2× TPU190 GB17.8 GB64.641551295.3 ms$1.50✓
4× TPU380 GB17.8 GB129.36206162.7 ms$0.7520✓
8× TPU760 GB17.8 GB258.62482396.3 ms$0.3760✓
Concurrency vs context length
On 1× TPU v5p (per-chip, 95GB) at FP16. KV cache scales linearly with context, so this is the number of users your GPUs can actually serve.
ContextConcurrent usersKV per requestFits?
4K344240.5 MB✓ fits
8K172481.0 MB✓ fits
33K431.88 GB✓ fits
66K213.76 GB✓ fits
131K107.52 GB✓ fits
Other GPUs that run Qwen 2.5-VL 7B (VLM)
Same workload, same quantization — ranked by value (tokens per dollar-hour).
GPUGPUsAggregate tok/sCost / 1M tok
RTX 3090 24GB1×521.5$0.1332
RTX 5090 32GB1×998.4$0.1391
RTX 4090 24GB1×561.6$0.1978
AMD MI300X 192GB1×2953$0.2352
V100 SXM2 32GB1×501.4$0.2770
A100 40GB SXM41×863.6$0.2895
Frequently asked questions

Can Qwen 2.5-VL 7B (VLM) run on TPU v5p (per-chip, 95GB)?

Yes. Qwen 2.5-VL 7B (VLM) at FP16 needs 17.8 GB including the KV cache, and TPU v5p (per-chip, 95GB) has 95 GB — so it fits on a single GPU.

How fast is Qwen 2.5-VL 7B (VLM) on TPU v5p (per-chip, 95GB)?

38.02 tokens/sec per stream and 456.3 tok/s aggregate with continuous batching at 8 concurrent users. Time to first token is 556.0 ms, and inter-token latency is 26.3 ms.

What does Qwen 2.5-VL 7B (VLM) on TPU v5p (per-chip, 95GB) cost per million tokens?

$2.56 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 Qwen 2.5-VL 7B (VLM) need on TPU v5p (per-chip, 95GB)?

14.0 GB for the weights at FP16, plus 3.8 GB of KV cache for 8 concurrent requests at 8K context — 17.8 GB total. Every extra concurrent user at long context costs more, so leave headroom.

Is TPU v5p (per-chip, 95GB) the right choice for Qwen 2.5-VL 7B (VLM)?

It works. Compare $2.56 per 1M tokens against the alternatives below — the cheapest option that fits the same workload is 2× RTX 3080 10GB at $0.0772, and the fastest is 1× Cerebras CS-3 (WSE-3) at 11143 tok/s.

Adjust the workload yourself

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

Open in calculator