# ds4

> Local DeepSeek 4 inference engine

antirez's engine to run DeepSeek 4 models on Metal, CUDA and ROCm.

- Source: https://github.com/antirez/ds4
- License: MIT
- Language: C
- Stars: 23203
- Forks: 2233
- Contributors: 55
- Last commit: 2026-09-20
- Purpose: AI & LLM tooling, Local inference
- Runs on: CLI
- For: Personal

## Worth score: 63/100

How much DigGitHub recommends it, from activity, adoption, docs, license and security signals.

- Popularity: 22/25
- Momentum: 10/20
- Maintenance: 15/25
- Community: 11/15
- Readiness: 5/15

## Alternatives

- [Ollama](https://diggithub.com/ollama/ollama.md): Run open-weight LLMs locally with one command
- [llama.cpp](https://diggithub.com/ggml-org/llama.cpp.md): LLM inference in C/C++ on CPUs and GPUs
- [Colibri](https://diggithub.com/JustVugg/colibri.md): Run large MoE models on everyday hardware
- [llmfit](https://diggithub.com/AlexsJones/llmfit.md): Find which LLMs run on your hardware
- [llamafile](https://diggithub.com/mozilla-ai/llamafile.md): Run an LLM from a single file
- [exo](https://diggithub.com/exo-explore/exo.md): Run frontier AI models across your own devices

---

Source page: https://diggithub.com/antirez/ds4
Updated: 2026-10-04