DeepSeek R1 Distill Qwen 32B
deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
32B distilled reasoning model, runs on a single 24GB consumer GPU (INT4).
Parameters
32.5B
Context
128K
Downloads
2.9M
Likes
6.4K
Architecture
transformer
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 81.5 GB RAM 122.3 · Disk 133.4 | 100.8 GB RAM 151.1 · Disk 159.4 | 201.1 GB RAM 301.7 · Disk 198.4 |
| FP88bit | 43.2 GB RAM 64.9 · Disk 86.2 | 60.7 GB RAM 91.1 · Disk 112.2 | 159.4 GB RAM 239.1 · Disk 151.2 |
| INT88bit | 43.2 GB RAM 64.9 · Disk 86.2 | 60.7 GB RAM 91.1 · Disk 112.2 | 159.4 GB RAM 239.1 · Disk 151.2 |
| AWQ4bit | 24.8 GB RAM 37.2 · Disk 63.5 | 41.5 GB RAM 62.2 · Disk 89.5 | 139.3 GB RAM 208.9 · Disk 128.5 |
| GPTQ4bit | 25.1 GB RAM 37.6 · Disk 63.8 | 41.7 GB RAM 62.6 · Disk 89.8 | 139.5 GB RAM 209.3 · Disk 128.8 |
| GGUF4bit | 25.5 GB RAM 38.3 · Disk 64.4 | 42.2 GB RAM 63.3 · Disk 90.4 | 140.1 GB RAM 210.1 · Disk 129.4 |
GPU Recommendations
1× H100 SXM
80 GB VRAM
≈ 213.4 tok/s
Effective production throughput ≈ 149.4 tok/s
score 0.6278
4× Tesla V100
64 GB VRAM
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.5711
8× H200 SXM
1128 GB VRAM
≈ 978.3 tok/s
Effective production throughput ≈ 684.8 tok/s
score 0.2197
1× RTX A6000
48 GB VRAM
≈ 48.9 tok/s
Effective production throughput ≈ 34.2 tok/s
score 0.6604
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
RTX A6000
48 GB VRAM
$226
/mo · vast
L40S
48 GB VRAM
$293
/mo · vast
RTX PRO 6000 WS
48 GB VRAM
$482
/mo · vast
RTX PRO 6000 S
48 GB VRAM
$634
/mo · vast
A100 80GB
80 GB VRAM
$876
/mo · vast