RUN THIS LLM
Search local LLM hardware requirements
DeepSeek V4 Pro
DeepSeek · 1.6T MoE · General
DeepSeek's V4 preview. 1.6T MoE with 49B active, hybrid CSA/HCA attention, 1M context, native FP4+FP8 mixed precision.
VRAM Requirements
| Quantization | VRAM |
|---|---|
| Q4_K_M (smallest) | 960 GB |
| Q8_0 (balanced) | 1760 GB |
| FP16 (full quality) | 3200 GB |
Specifications
- Parameters: 1.6T MoE (49B active per token)
- Category: General
- Max context: 1M tokens
- System RAM: 1024 GB minimum
- HuggingFace: deepseek-ai/DeepSeek-V4-Pro
Benchmarks
- HumanEval: 94 — Code generation (pass@1)
- MATH: 90 — Competition-level math reasoning
Loading interactive analysis...