Qwen2.5-VL 72B
Qwen/Qwen2.5-VL-72B-Instruct
Parameters
72.7B
Context
32K
Downloads
528.9K
Likes
647
Architecture
transformer
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 174.9 GB RAM 262.3 · Disk 250.3 | 184.8 GB RAM 277.3 · Disk 276.3 | 208.5 GB RAM 312.7 · Disk 315.3 |
| FP88bit | 89.2 GB RAM 133.8 · Disk 144.6 | 95.3 GB RAM 142.9 · Disk 170.6 | 115.0 GB RAM 172.6 · Disk 209.6 |
| INT88bit | 89.2 GB RAM 133.8 · Disk 144.6 | 95.3 GB RAM 142.9 · Disk 170.6 | 115.0 GB RAM 172.6 · Disk 209.6 |
| AWQ4bit | 48.0 GB RAM 72.0 · Disk 93.8 | 52.2 GB RAM 78.3 · Disk 119.8 | 70.1 GB RAM 105.1 · Disk 158.8 |
| GPTQ4bit | 48.5 GB RAM 72.8 · Disk 94.5 | 52.8 GB RAM 79.1 · Disk 120.5 | 70.7 GB RAM 106.0 · Disk 159.5 |
| GGUF4bit | 49.6 GB RAM 74.4 · Disk 95.8 | 53.9 GB RAM 80.8 · Disk 121.8 | 71.8 GB RAM 107.7 · Disk 160.8 |
GPU Recommendations
1× H100 SXM
80 GB VRAM
≈ 95.4 tok/s
Effective production throughput ≈ 66.8 tok/s
score 0.6587
4× Tesla V100
64 GB VRAM
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.5769
8× H200 SXM
1128 GB VRAM
≈ 437.4 tok/s
Effective production throughput ≈ 306.2 tok/s
score 0.2197
1× A100 SXM4
80 GB VRAM
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.6569
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
A100 SXM4
80 GB VRAM
$343
/mo · vast
A100 PCIE
80 GB VRAM
$523
/mo · vast
A100 80GB
80 GB VRAM
$876
/mo · vast
H100 SXM
80 GB VRAM
$1,404
/mo · vast
H200 NVL
141 GB VRAM
$1,610
/mo · vast