DeepSeek-Coder-V2 Lite 16B
deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct
Parameters
16B
Context
128K
Downloads
545.5K
Likes
630
Architecture
moe_transformer
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 42.7 GB RAM 64.0 · Disk 85.5 | 60.1 GB RAM 90.2 · Disk 111.5 | 158.7 GB RAM 238.1 · Disk 150.5 |
| FP88bit | 23.8 GB RAM 35.7 · Disk 62.3 | 40.4 GB RAM 60.6 · Disk 88.3 | 138.2 GB RAM 207.3 · Disk 127.3 |
| INT88bit | 23.8 GB RAM 35.7 · Disk 62.3 | 40.4 GB RAM 60.6 · Disk 88.3 | 138.2 GB RAM 207.3 · Disk 127.3 |
| AWQ4bit | 14.7 GB RAM 32.0 · Disk 51.1 | 30.9 GB RAM 46.4 · Disk 77.1 | 128.3 GB RAM 192.4 · Disk 116.1 |
| GPTQ4bit | 14.9 GB RAM 32.0 · Disk 51.2 | 31.1 GB RAM 46.6 · Disk 77.2 | 128.4 GB RAM 192.6 · Disk 116.2 |
| GGUF4bit | 15.1 GB RAM 32.0 · Disk 51.5 | 31.3 GB RAM 46.9 · Disk 77.5 | 128.7 GB RAM 193.0 · Disk 116.5 |
GPU Recommendations
1× H100 SXM
80 GB VRAM
≈ 433.4 tok/s
Effective production throughput ≈ 303.4 tok/s
score 0.5926
2× Tesla V100
32 GB VRAM
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.5933
8× H200 SXM
1128 GB VRAM
≈ 1987.1 tok/s
Effective production throughput ≈ 1391 tok/s
score 0.2197
1× RTX A6000
48 GB VRAM
≈ 99.4 tok/s
Effective production throughput ≈ 69.6 tok/s
score 0.6629
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
RTX A6000
48 GB VRAM
$210
/mo · vast
RTX 5090
32 GB VRAM
$244
/mo · vast
L40S
48 GB VRAM
$341
/mo · vast
RTX PRO 6000 WS
48 GB VRAM
$482
/mo · vast
RTX PRO 6000 S
48 GB VRAM
$633
/mo · vast