Run NVIDIA-Nemotron-3.5-Transcribe-30B-A4B-FP8 on a cloud GPU
NVIDIA-Nemotron-3.5-Transcribe-30B-A4B-FP8 is a Speech to text model by nvidia. See how much GPU memory it needs, which GPUs it fits on today and what it costs to run it yourself, billed per second.
GPUs it fits on today
Our estimate is about 42 GB for inference with F8_E4M3 weights, including working memory. Real use depends on resolution, batch size and context length.
| GPU | Memory | Per hour | 10 hours | |
|---|---|---|---|---|
| Q RTX 8000 | 45 GB | $0.54 | $5.40 | Rent |
| A100 SXM4 | 80 GB | $0.75 | $7.50 | Rent |
| A40 | 45 GB | $0.87 | $8.70 | Rent |
| CMP 170HX | 64 GB | $0.88 | $8.80 | Rent |
| RTX A6000 | 48 GB | $0.94 | $9.40 | Rent |
Live starting prices from our catalog. You pay per second while the machine runs.
How to run NVIDIA-Nemotron-3.5-Transcribe-30B-A4B-FP8
- 01
Rent a GPU
Pick one of the GPUs above. Billing is per second; stop the machine and you stop paying.
- 02
Open a terminal or Jupyter
Connect over SSH or open Jupyter in your browser. PyTorch and CUDA are ready to use.
- 03
Download and run the model
Run the commands below. The weights download straight from Hugging Face to your machine.
pip install -U transformers accelerate
python -c "from transformers import pipeline; p = pipeline('automatic-speech-recognition', model='nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B-FP8', device=0); print(p)"Why run it yourself?
Transcribe hours of audio for the cost of the GPU minutes it takes. Recordings never leave your machine.
Questions
- How much GPU memory does NVIDIA-Nemotron-3.5-Transcribe-30B-A4B-FP8 need?
- By our estimate about 42 GB for inference (34B parameters). Higher resolution, longer context or bigger batches need more.
- How much does it cost to run NVIDIA-Nemotron-3.5-Transcribe-30B-A4B-FP8?
- From $0.54 per hour on Q RTX 8000, billed per second. Ten hours cost about $5.40.
- Do I need a subscription?
- No. You pay only for the time your machine runs, from your USD balance. Stop it whenever you are done.
Try it on your own GPU in minutes
Pick a GPU, open Jupyter or SSH and run any open model. You pay per second and can stop anytime; no subscription, no setup fee.
- Per-second billing in USD
- PyTorch and CUDA ready in minutes
- Rent from ChatGPT or Claude too