RUN THIS LLM
Search local LLM hardware requirements
DeepSeek Coder V2 Lite
DeepSeek · 16B MoE · Code
Lightweight MoE code model. Only 2.4B active params but strong coding performance.
VRAM Requirements
| Quantization | VRAM |
|---|---|
| Q4_K_M (smallest) | 9.6 GB |
| Q8_0 (balanced) | 17.6 GB |
| FP16 (full quality) | 32 GB |
Specifications
- Parameters: 16B MoE (2.4B active per token)
- Category: Code
- Max context: 128K tokens
- System RAM: 16 GB minimum
- HuggingFace: deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct
Benchmarks
- HumanEval: 72 — Code generation (pass@1)
Loading interactive analysis...