Local LLM model page

DeepSeek V4 Flash (284B MoE)

Efficient DeepSeek V4 variant: 284B total, 13B active, 1M-token context. Flash-Max can approach Pro reasoning with larger thinking budget. MIT licensed.

Find the best model for my hardwareBrowse all 183 LLMs
Parameters
284B (13B active)
Minimum RAM
256 GB
Model size
170 GB
Quantization
FP4/FP8

Can DeepSeek V4 Flash (284B MoE) run locally?

DeepSeek V4 Flash (284B MoE) is best suited for server-grade or multi-GPU systems. LocalClaw recommends FP4/FP8 as the default quantization, with at least 256 GB RAM.

Search term for LM Studio or compatible runtimes: deepseek-v4-flash

Hugging Face repository: deepseek-ai/DeepSeek-V4-Flash

chatcodereasoningpoweragenticlong-contextgeneral

Strengths

  • Efficient DeepSeek V4 variant: 284B total, 13B active, 1M-token context. Flash-Max can approach Pro reasoning with larger thinking budget. MIT licensed.

Limitations

  • Performance depends heavily on quantization, RAM bandwidth and runtime support.

Best use cases

  • chat
  • code
  • reasoning
  • power
  • agentic
  • long-context
  • general

Benchmarks

Speed: 5/10

Quality: 9/10

Coding: 9/10

Reasoning: 9/10

Technical details

Developer: deepseek-flash

License: See model repository

Context window: Unknown tokens

Architecture: See model card

Released: 2026-05