KwaiVGI
Kuaishou/KwaiVGI
Kuaishou KwaiVGI video generation model.
Parameters
4B
Context
16K
Downloads
80K
Likes
500
Architecture
kwai
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 20.3 GB RAM 32.0 · Disk 50.6 | 24.6 GB RAM 36.9 · Disk 76.6 | 32.9 GB RAM 49.3 · Disk 115.6 |
| FP88bit | 12.9 GB RAM 32.0 · Disk 44.8 | 16.9 GB RAM 32.0 · Disk 70.8 | 24.9 GB RAM 37.3 · Disk 109.8 |
| INT88bit | 12.9 GB RAM 32.0 · Disk 44.8 | 16.9 GB RAM 32.0 · Disk 70.8 | 24.9 GB RAM 37.3 · Disk 109.8 |
| AWQ4bit | 9.3 GB RAM 32.0 · Disk 42.0 | 13.2 GB RAM 32.0 · Disk 68.0 | 21.0 GB RAM 32.0 · Disk 107.0 |
| GPTQ4bit | 9.4 GB RAM 32.0 · Disk 42.1 | 13.3 GB RAM 32.0 · Disk 68.1 | 21.0 GB RAM 32.0 · Disk 107.1 |
| GGUF4bit | 9.5 GB RAM 32.0 · Disk 42.1 | 13.3 GB RAM 32.0 · Disk 68.1 | 21.1 GB RAM 32.0 · Disk 107.1 |
GPU Recommendations
1× RTX 5090
32 GB VRAM
≈ 288 video-sec/GPU-h
Effective production throughput ≈ 172.8 video-sec/GPU-h
score 0.6074
1× RTX 4070S Ti
16 GB VRAM
≈ 288 video-sec/GPU-h
Effective production throughput ≈ 172.8 video-sec/GPU-h
score 0.669
8× RTX 5090
256 GB VRAM
≈ 288 video-sec/GPU-h
Effective production throughput ≈ 172.8 video-sec/GPU-h
score 0.3581
1× RTX 3090
24 GB VRAM
≈ 288 video-sec/GPU-h
Effective production throughput ≈ 172.8 video-sec/GPU-h
score 0.6418
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
RTX 4070S Ti
16 GB VRAM
$49.5
/mo · vast
RTX 4080S
16 GB VRAM
$50.0
/mo · vast
RTX 3090
24 GB VRAM
$76.5
/mo · vast
RTX 5070 Ti
16 GB VRAM
$86.3
/mo · vast
Tesla V100
16 GB VRAM
$88.4
/mo · vast