FishFace comments on Local LLM agents

submitted 5 months ago by Kkk2237pl@lemmy.world to c/programming@programming.dev

30 comments fedilink hide all child comments

Has anyone tried in organization to use self hosted llm models for agentic programming?

Im curious if it makes any sense. My organization spends fortune on tokens from us companies. I want to recommend something…

[–] FishFace@piefed.social 9 points 5 months ago (6 children)

As far as I understand, the only way to get anything resembling usable output for coding is with massive, expensive, labouriously hand-tuned models, not local ones.

permalink

fedilink

source

hideshow 6 child comments

[–] Jestzer@lemmy.world 6 points 5 months ago

^^^ This. Tragically, locally run LLMs don’t even hold a candle to “good” cloud-based LLMs like Claude Code.

permalink

fedilink

source

parent

[–] locuester@lemmy.zip 5 points 5 months ago

Qwen 3.6 27B dense is really good. Very usable coding output

permalink

fedilink

source

parent

[–] Kkk2237pl@lemmy.world [S] 3 points 5 months ago (2 children)

I see that qwen 3.5 has pretty good performance and can be run on macbook with 64GB ram

permalink

fedilink

source

parent

hideshow 2 child comments

[–] Penta@lemmy.world 10 points 5 months ago

Qwen 3.6 is even better

permalink

fedilink

source

parent

[–] SmoothLiquidation@lemmy.world 1 point 5 months ago

I have played with qwen3-coder:30b for my hobby stuff running on my M5 max MacBook and it does alright. It is fast enough and I used ollama tools to let it request files. I haven’t used anything like Claude code to compare it to though, only a bit of the ChatGPT free tier stuff.