A30 vs V100 32GB
Both run on the same nodes, the same templates and the same tiers here — so the choice comes down to memory, bandwidth and price. Every figure below is from the catalogue or the vendor datasheet.
- A30
- $0.05/h24 GB · Ampere · 32 available
- V100
- $0.06/h32 GB · Volta · 32 available
- Price gap
- 17%On spot, per GPU-hour
- Over 100 h
- $1.00Difference for the same GPU-hours
What actually separates them.
Four differences that change a decision, computed from the numbers rather than asserted.
Bigger models fit on the V100
32 GB against 24 GB. At FP16 that is about 12B parameters on one GPU, against 9B.
The A30 costs 17% less per hour
$0.05 against $0.06 on spot — $1.00 of difference over 100 GPU-hours.
Token generation is faster on the A30
933 GB/s against 900 GB/s. Inference reads the whole model from memory for every token, so bandwidth, not FLOPS, usually sets the ceiling.
More memory per dollar on the V100
533 GB per dollar per hour against 480 GB. The usual tie-breaker when both cards fit the job.
Every figure we hold on both cards.
A row is marked only when both values exist and one is genuinely better. A30 leads on 6 rows, V100 on 4.
| A30 | V100 32GB | |
|---|---|---|
| GPU memoryDecides which models fit at all | 24 GB HBM2 | 32 GB HBM2 |
| Memory bandwidthSets token throughput on inference | 933 GB/s | 900 GB/s |
| FP16 tensor | 165 TFLOPS | 125 TFLOPS |
| FP32 | 10 TFLOPS | 16 TFLOPS |
| 8-bit float (FP8)Halves the memory a model needs, where the stack supports it | Not supported | Not supported |
| Architecture | Ampere (2021) | Volta (2018) |
| Board powerLower is cheaper to run at scale — for us, not for your bill | 165 W | 300 W |
| GPU-to-GPU linkOnly matters for multi-GPU training | PCIe only | NVLink |
| Node sizes | 1×, 2×, 4× | 1×, 2×, 4× |
| Spot price | $0.05/h | $0.06/h |
| On-demand price | $0.08/h | $0.09/h |
| Memory per dollarGPU memory per $1 of spot time, per hour | 480 GB | 533 GB |
| 30-day reclaim rateShare of spot instances reclaimed | <5% | <5% |
| Available now | 32 GPUs | 32 GPUs |
Specifications come from each vendor's datasheet — linked on the A30 page and the V100 page. Prices are ours, per GPU-hour, billed per minute.
The same models, on one GPU of each.
Weights only, with 20% of the memory kept free for activations and the KV cache. A cross means one GPU is not enough — the model page has the multi-GPU arithmetic.
| Model | Precision | A30 24 GB | V100 32 GB |
|---|---|---|---|
| Llama 3.1 8B8B parameters | FP16 · 16 GB | Fits | Fits |
| INT4 · 4 GB | Fits | Fits | |
| Qwen2.5 32B32B parameters | FP16 · 64 GB | Needs 4× node | Needs 4× node |
| INT4 · 16 GB | Fits | Fits | |
| Llama 3.3 70B70B parameters | FP16 · 140 GB | Needs more than one node | Needs more than one node |
| INT4 · 35 GB | Needs 2× node | Needs 2× node | |
| Mixtral 8x22B141B parameters | FP16 · 282 GB | Needs more than one node | Needs more than one node |
| INT4 · 71 GB | Needs 4× node | Needs 4× node |
The same 100 GPU-hours on each.
Compute only, billed per minute. Storage is $0.08 per GB-month on either card.
| A30 | V100 | |
|---|---|---|
| One hour, spot | $0.05 | $0.06 |
| One day, spot | $1.20 | $1.44 |
| 100 GPU-hours, spot | $5.00 | $6.00 |
| 100 GPU-hours, on-demand | $8.00 | $9.00 |
| A month, reserved | $43.80 | $43.80 |
Take the A30 when…
- Fine-tuning small models — LoRA on 7B–13B models, and full fine-tuning below 3B.
- Image and video generation — Diffusion pipelines run entirely in memory at this capacity.
- Partitioned serving (MIG) — The GPU can be split into isolated instances, each with its own memory and compute slice.
Take the V100 when…
- Fine-tuning small models — LoRA on 7B–13B models, and full fine-tuning below 3B.
- Image and video generation — Diffusion pipelines run entirely in memory at this capacity.
- Multi-GPU jobs over NVLink — GPU-to-GPU traffic stays off the PCIe bus, which is what makes tensor and pipeline parallelism worth it.
A30 or V100?
Which is cheaper, the A30 or the V100?
The A30, at $0.05 per GPU-hour on spot against $0.06 — 17% less. On-demand: $0.08 against $0.09.
Which one should I take for inference?
Take the one that holds the model with room for the KV cache, then the one with more memory bandwidth. Here that is the V100 on capacity (32 GB) and the A30 on bandwidth.
Can I run both on the same job?
Not inside one instance — a node holds GPUs of a single model. You can run two instances in parallel, on the same persistent volume in turn, or split the work: a big card for the model that needs capacity, a cheap one for the parts that do not.
Other comparisons: the full list · all 47 models
Try both for less than the price of deciding.
An hour on each costs $0.11 on spot. Nothing renews, and your disk moves between them.
Billed per minute from the moment the instance is reachable. Minimum credit $40, no subscription.