Qwen/Qwen3-VL-8B-Instruct
Qwen/Qwen3-VL-8B-Instruct
Parameters
8.3B
Context
256K
Downloads
17.1M
Likes
1.2K
Architecture
qwen3_vl
36 layers · 8 KV heads · head 128 · hidden 4096
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 23.2 GB RAM 34.8 · Disk 63.0 | 26.5 GB RAM 39.8 · Disk 89.0 | 45.2 GB RAM 67.8 · Disk 128.0 |
| FP88bit | 13.5 GB RAM 32.0 · Disk 51.0 | 16.4 GB RAM 32.0 · Disk 77.0 | 34.6 GB RAM 51.9 · Disk 116.0 |
| INT88bit | 13.5 GB RAM 32.0 · Disk 51.0 | 16.4 GB RAM 32.0 · Disk 77.0 | 34.6 GB RAM 51.9 · Disk 116.0 |
| AWQ4bit | 8.8 GB RAM 32.0 · Disk 45.2 | 11.5 GB RAM 32.0 · Disk 71.2 | 29.5 GB RAM 44.3 · Disk 110.2 |
| GPTQ4bit | 8.9 GB RAM 32.0 · Disk 45.3 | 11.5 GB RAM 32.0 · Disk 71.3 | 29.6 GB RAM 44.4 · Disk 110.3 |
| GGUF4bit | 9.0 GB RAM 32.0 · Disk 45.4 | 11.6 GB RAM 32.0 · Disk 71.4 | 29.7 GB RAM 44.6 · Disk 110.4 |
GPU Recommendations
1× RTX 5080
16 GB VRAM
≈ 240.6 tok/s
Effective production throughput ≈ 168.4 tok/s
score 0.6838
1× Tesla V100
16 GB VRAM
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.6798
8× H100 SXM
640 GB VRAM
≈ 2686.7 tok/s
Effective production throughput ≈ 1880.7 tok/s
score 0.2026
1× RTX 3080 Ti
12 GB VRAM
≈ 228.6 tok/s
Effective production throughput ≈ 160 tok/s
score 0.6415
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
Tesla V100
16 GB VRAM
$50.3
/mo · vast
RTX 3080 Ti
12 GB VRAM
$79.5
/mo · vast
RTX 3090
24 GB VRAM
$89.2
/mo · vast
RTX 5070 Ti
16 GB VRAM
$89.2
/mo · vast
RTX 5070
12 GB VRAM
$138
/mo · vast