Qwen/Qwen3-14B
Qwen/Qwen3-14B
Efficient 14B model, runs on single 16-24GB consumer GPU.
Parameters
13.4B
Context
40K
Downloads
13M
Likes
541
Architecture
qwen3
40 layers · 8 KV heads · head 128 · hidden 5120
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 39.8 GB RAM 59.7 · Disk 82.0 | 57.2 GB RAM 85.7 · Disk 108.0 | 155.6 GB RAM 233.5 · Disk 147.0 |
| FP88bit | 22.4 GB RAM 33.6 · Disk 60.5 | 38.9 GB RAM 58.4 · Disk 86.5 | 136.6 GB RAM 204.9 · Disk 125.5 |
| INT88bit | 22.4 GB RAM 33.6 · Disk 60.5 | 38.9 GB RAM 58.4 · Disk 86.5 | 136.6 GB RAM 204.9 · Disk 125.5 |
| AWQ4bit | 14.0 GB RAM 32.0 · Disk 50.2 | 30.2 GB RAM 45.2 · Disk 76.2 | 127.5 GB RAM 191.2 · Disk 115.2 |
| GPTQ4bit | 14.1 GB RAM 32.0 · Disk 50.3 | 30.3 GB RAM 45.4 · Disk 76.3 | 127.6 GB RAM 191.4 · Disk 115.3 |
| GGUF4bit | 14.3 GB RAM 32.0 · Disk 50.6 | 30.5 GB RAM 45.8 · Disk 76.6 | 127.8 GB RAM 191.7 · Disk 115.6 |
GPU Recommendations
1× H100 SXM
80 GB VRAM
≈ 468.5 tok/s
Effective production throughput ≈ 328 tok/s
score 0.5881
2× Tesla V100
32 GB VRAM
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.5947
8× H200 SXM
1128 GB VRAM
≈ 2148.3 tok/s
Effective production throughput ≈ 1503.8 tok/s
score 0.2197
1× RTX 5090
32 GB VRAM
≈ 250.6 tok/s
Effective production throughput ≈ 175.4 tok/s
score 0.6482
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
RTX 5090
32 GB VRAM
$155
/mo · vast
A100 SXM4
80 GB VRAM
$225
/mo · vast
RTX A6000
48 GB VRAM
$293
/mo · vast
L40S
48 GB VRAM
$314
/mo · vast
A100 PCIE
80 GB VRAM
$355
/mo · vast