# Python or Rust? Coding with local LLMs — retained record

> Retained reddit record from reddit:LocalLLaMA.

Canonical: https://highsignal.app/data/records/1046cb9b823f2a20

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:LocalLLaMA)  published 2026-10-21 14:09 ingested 2026-10-22 02:43

[original source ↗](https://www.reddit.com/r/LocalLLaMA/comments/1xbo6if/python_or_rust_coding_with_local_llms/)

## retained text



```
I've noticed that a lot of software is being cloned using Rust nowadays, and I was wondering why not use it with local LLMs for simple tasks too.

Do you think it's better to use Rust or Python when coding with a local LLM?

Nowadays everyone is vibe coding, but local LLMs are generally not that good, especially smaller models running on low-budget hardware.

I usually code in Python, but sometimes I feel like the model doesn't catch bugs, so I was wondering if Rust and its compiler could help avoid some of these issues.

I'd also be interested in hearing your experiences with different context window sizes and models running on 12 or 24 GB of VRAM, with 64 or 128 GB of system RAM.

Do you think Rust's compiler feedback is worth the extra complexity for smaller local models, or is Python still the better choice?
```



document fetched 2026-10-22 00:44 · canonical https://reddit.com/r/LocalLLaMA/comments/1xbo6if/python_or_rust_coding_with_local_llms
