Which GPU runs which model · DeepSeek
Which GPU runs DeepSeek-R1-Distill-Qwen-32B, and what it costs per hour
DeepSeek-R1-Distill-Qwen-32B (32B) takes about 64 GB of VRAM as FP16 weights and about 19 to 20 GB at Q4_K_M, before the KV cache for your context window. The table lists the cards on cloudgpu.app by VRAM class, whether the model fits, and the price you will actually be billed, read on 2026-09-15.
Recommended card: RTX 5090 32G (32 GB) at $0.59 per hour, the lowest-priced card in a VRAM class that fits.
| VRAM | Cards | Price per hour | Fit |
|---|---|---|---|
| 24 GB | RTX 3090 24G · RTX 4090D 24G · RTX 4090 24G | $0.21 – $0.33/h | Tight fit Q4 fits with about 8k context |
| 32 GB | RTX 5090 32G | $0.59/h | Fits Q4 comfortable with 16k to 32k context |
| 48 GB | RTX A6000 48G · RTX 4090 48G | $0.44 – $0.65/h | Fits Q4 long context or Q8 |
| 80 GB | A800 80G · A100 80G | $1.09 – $1.49/h | Fits FP16 fits |
| 96 GB | H20 96G · RTX PRO 6000 96G | $1.29 – $1.49/h | Fits FP16 fits |
Prices are the retail rates on cloudgpu.app read 2026-09-15 and equal what the billing system charges; VRAM figures were checked on 2026-09-08. Stock changes during the day.
Reading the table
"Fits" is judged for the Q4_K_M weights plus a working context window; FP16 needs roughly twice the VRAM. The note under each verdict says how much context you get.
Data-centre cards (80 GB and 96 GB) are not cheap on cloudgpu.app. If the model fits on a consumer card, the consumer card is the better deal. The prices here are the same numbers the billing system charges.
Questions about running DeepSeek-R1-Distill-Qwen-32B
How much VRAM does DeepSeek-R1-Distill-Qwen-32B need?
DeepSeek-R1-Distill-Qwen-32B has 32B parameters. The FP16 weights take about 64 GB; at Q4_K_M the weights take about 19 to 20 GB. Add 1 to 8 GB for the KV cache depending on context length and batch size.
What is the cheapest GPU that runs DeepSeek-R1-Distill-Qwen-32B?
On cloudgpu.app the lowest-priced card we recommend for DeepSeek-R1-Distill-Qwen-32B is the RTX 5090 32G with 32 GB at $0.59 per hour (Q4 comfortable with 16k to 32k context). Prices read 2026-09-15; the pricing page shows live stock.
Can I run DeepSeek-R1-Distill-Qwen-32B on an RTX 4090 with 24 GB?
Yes, but it is a tight fit. On a 24 GB card: Q4 fits with about 8k context. For real work a 32 GB RTX 5090 ($0.59 per hour) or a 48 GB card is the practical minimum.
What does it cost per hour to run DeepSeek-R1-Distill-Qwen-32B on cloudgpu.app?
From $0.59 per hour on the RTX 5090 32G, billed per minute. A 20-minute test is about $0.20; one hour is pre-authorised when the instance starts and unused minutes are refunded on stop. Cards in higher VRAM classes cost more; the table above lists each.
How do I start DeepSeek-R1-Distill-Qwen-32B on a rented GPU?
Open Rent a GPU, choose the DeepSeek-R1 32B template and a card from the table above. The instance is ready in about 100 seconds with a public HTTPS endpoint, and the template already includes DeepSeek-R1-Distill-Qwen-32B.
Do I need a credit card or a Chinese account?
No Chinese account, phone number or ID is required. You can pay by card (Visa or Mastercard, any amount from $5) or USDT (TRC-20). A first top-up of $10 or more gets a $5 bonus, so a $10 card payment starts you with $15 of GPU time.