Selecting Local Agent Models for 36GB VRAM

Developers are discussing the best local LLMs to run coding and general agents within a 36GB VRAM limit. Candidates like Qwen3.6 27B are being evaluated for their ability to handle continuous software projects effectively.

2026-07-28 ~ 2026-07-28 · 2 related posts

• Which local LLM is best for a 36GB VRAM coding agent setup? — hovikyan · 2026-07-28
• Which local LLM should run a long-lived agent on 36GB of VRAM? — hovikyan · 2026-07-28