GPU class · 48 GB
A6000
What does A6000 cost and what fits on it?
A6000 has 48 GB and costs $0.65 per hour to train on, $0.75 to serve. 36 models in the catalogue fit half-precision LoRA on it and 5 more fit in four-bit.
The most plentiful 48GB datacentre card, which in practice means it is the 48GB class least likely to keep you waiting for capacity. Performance sits between the A40 and the RTX 6000 Ada, and so does the price.
When to pick this class
The most plentiful 48GB datacentre card, and in practice that means it is the 48GB class least likely to keep you waiting for capacity. When a run is queued behind demand elsewhere, availability is worth more than a few cents an hour.
Performance sits between the A40 and the RTX 6000 Ada, and so does the price, which makes it the sensible default at this capacity when you have no specific reason to prefer either of the others.
Rates
| VRAM | 48 GB |
|---|---|
| Training | $0.65 / GPU-hourMetered per GPU-second, one-minute floor |
| Serving | $0.75 / GPU-hourWhile a replica is resident |
| Warm for a day | $1824 hours resident, regardless of traffic |
| Warm for 30 days | $540Set the endpoint to scale to zero if nobody is waiting |
| Models — LoRA | 36 |
| Models — QLoRA only | 5 |
Largest models that fit for LoRA
Half precision, frozen base, within 48 GB.
Mistral Small 3 24B
24BNeeds 48 GB · Apache 2.0
LFM2 24B-A2B
24BNeeds 48 GB · LFM Open
gpt-oss 20B
20.9BNeeds 40 GB · Apache 2.0
Qwen3 14B
14BNeeds 28 GB · Apache 2.0
Phi-4 14B
14BNeeds 28 GB · MIT
Mistral Nemo 12B
12BNeeds 24 GB · Apache 2.0
Gemma 3 12B
12BNeeds 24 GB · Gemma Terms of Use
Falcon 3 10B
10BNeeds 22 GB · TII Falcon LLM
Models that need four-bit training here
These exceed 48 GB in half precision but fit quantised. Quantised training is slower per step, so compare the total run cost against the next class up rather than the hourly rate alone.
Compared with its neighbours
The classes immediately either side of A6000 on capacity and rate.
| Class | VRAM | Training | Serving | Versus this one |
|---|---|---|---|---|
| A40 | 48 GB | $0.42 | $0.48 | same capacity, $0.23/hr cheaper |
| RTX 6000 Ada | 48 GB | $0.61 | $0.71 | same capacity, $0.04/hr cheaper |
| L40S | 48 GB | $0.78 | $0.90 | same capacity, $0.13/hr dearer |
| A100 80 GB | 80 GB | $1.68 | $1.94 | 32 GB more, $1.03/hr dearer |
Frequently asked questions
How much does A6000 cost per hour?
$0.65 per GPU-hour for training and $0.75 for serving, metered per GPU-second above a one-minute floor. A continuously warm endpoint on this class is about $18 a day, or $540 over thirty days.
What models fit on A6000?
36 models in the catalogue fit half-precision LoRA training within 48 GB, and 5 more fit in four-bit. 37 can be served from this class before accounting for the attention cache.
Is A6000 the cheapest option?
Cheapest per hour is not the same as cheapest per run. Total cost is the rate multiplied by wall time, so a faster class that finishes sooner often costs less — particularly when the cheaper alternative would require quantised training, which is slower per step.
Last verified 6 August 2026.
Start with the free tier
A magic link creates your account, your tenant and your first API key. No card until you ask for compute.