← Tool Radar
library

antirez/ds4

A specialized inference engine for DeepSeek 4 Flash on Metal/CUDA, built by a renowned engineer, but its narrow scope limits broader appeal.

6.1Overall
Utility6
Onboarding4
Craft8
Niche fit6
Longevity6

Five dimensions scored independently; overall is a weighted average. Scores are only comparable within this same rubric.

Good for

Researchers or developers who specifically need to run DeepSeek 4 Flash efficiently on local Metal or CUDA hardware.

Not for

Users looking for a general-purpose local LLM inference solution, those unwilling to compile C code, or those without a specific need for the DeepSeek 4 Flash model.

Project description

DeepSeek 4 Flash local inference engine for Metal and CUDA

Alternatives

llama.cppMLC LLMOllama
Visit site1,100 · Stars at evalEvaluated 2026-05-14

Similar tools