Best Local LLM for Coding: 8GB to 24GB VRAM Picks

Description and use

Step‑by‑step recipe for selecting and running a local coding LLM based on GPU VRAM using Ollama.

Problem

Developers need a fast, offline coding assistant that fits their GPU VRAM constraints.

Outcome

Users can run a local coding LLM that matches their GPU VRAM, achieving fast autocomplete and refactor capabilities without external API costs.

Workflow

• 1. Determine GPU VRAM and choose model from table
• 2. Install Ollama (Linux) with curl script
• 3. Pull the selected model with ollama pull
• 4. Run interactively with ollama run or expose API via curl
• 5. Export OPENAI_BASE_URL for tools that expect it

Evidence

The best local LLM for coding right now is Qwen3 Coder 30B A3B on a 24 GB card, Qwen2.5 Coder 14B at Q4 on 12–16 GB, and Qwen2.5 Coder 7B at Q4 on 8 GB.

Source

Five commands, from nothing to an OpenAI-compatible API your editor can use: ...

Source