OPEN MODELS / Vision language models
Run DeepSeek-V4.1-Flash-1017-oQ4e-ngram-mtp on a cloud GPU
DeepSeek-V4.1-Flash-1017-oQ4e-ngram-mtp is a Vision language models model by mlx-community. 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 348 GB for inference with F8_E4M3 weights, including working memory. Real use depends on resolution, batch size and context length. With 4-bit quantization it needs about 209 GB.
| GPU | Memory | Per hour | 10 hours | |
|---|---|---|---|---|
| 8× Q RTX 8000 | 360 GB | $4.24 | $42.40 | Rent |
| 4× RTX PRO 6000 WS | 384 GB | $8.08 | $80.80 | Rent |
| 4× RTX PRO 6000 Max-Q | 384 GB | $9.12 | $91.20 | Rent |
| 8× L40S | 384 GB | $9.20 | $92.00 | Rent |
| 4× RTX PRO 6000 S | 384 GB | $10.08 | $100.80 | Rent |
Live starting prices from our catalog. You pay per second while the machine runs.
How to run DeepSeek-V4.1-Flash-1017-oQ4e-ngram-mtp
- 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 vllm
vllm serve mlx-community/DeepSeek-V4.1-Flash-1017-oQ4e-ngram-mtp --host 0.0.0.0 --port 8000Why run it yourself?
Run document and image understanding on your own GPU: no per-image fee, and files stay on your machine.
Questions
- How much GPU memory does DeepSeek-V4.1-Flash-1017-oQ4e-ngram-mtp need?
- By our estimate about 348 GB for inference (289B parameters). Higher resolution, longer context or bigger batches need more.
- How much does it cost to run DeepSeek-V4.1-Flash-1017-oQ4e-ngram-mtp?
- From $4.24 per hour on 8× Q RTX 8000, billed per second. Ten hours cost about $42.40.
- 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