Qwen3.5 122B-A10B
Qwen/Qwen3.5-122B-A10B
Parameters
122B
Context
256K
Downloads
2.1M
Likes
605
Architecture
moe_transformer
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 292.4 GB RAM 438.6 · Disk 393.5 | 321.2 GB RAM 481.8 · Disk 419.5 | 431.2 GB RAM 646.8 · Disk 458.5 |
| FP88bit | 148.7 GB RAM 223.0 · Disk 216.3 | 171.0 GB RAM 256.5 · Disk 242.3 | 274.4 GB RAM 411.6 · Disk 281.3 |
| INT88bit | 148.7 GB RAM 223.0 · Disk 216.3 | 171.0 GB RAM 256.5 · Disk 242.3 | 274.4 GB RAM 411.6 · Disk 281.3 |
| AWQ4bit | 79.5 GB RAM 119.3 · Disk 131.0 | 98.7 GB RAM 148.0 · Disk 157.0 | 198.9 GB RAM 298.4 · Disk 196.0 |
| GPTQ4bit | 80.4 GB RAM 120.6 · Disk 132.1 | 99.6 GB RAM 149.4 · Disk 158.1 | 199.9 GB RAM 299.9 · Disk 197.1 |
| GGUF4bit | 82.2 GB RAM 123.3 · Disk 134.3 | 101.5 GB RAM 152.2 · Disk 160.3 | 201.9 GB RAM 302.8 · Disk 199.3 |
GPU Recommendations
Best ValueBEST
4× H200 SXM
564 GB VRAM
$10,220/mo
≈ 195.5 tok/s
Effective production throughput ≈ 136.9 tok/s
score 0.316
Cheapest
8× Tesla V100
128 GB VRAM
$169/mo
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.551
Performance
8× H200 SXM
1128 GB VRAM
$20,440/mo
≈ 260.6 tok/s
Effective production throughput ≈ 182.4 tok/s
score 0.2197
Min Complexity
1× H200 SXM
141 GB VRAM
$2,555/mo
≈ 81.4 tok/s
Effective production throughput ≈ 57 tok/s
score 0.6576
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
H200 SXM
141 GB VRAM
$2,555
/mo · lambda
B200
180 GB VRAM
$3,013
/mo · vast
B200 SXM
180 GB VRAM
—