# QDA: Closed-Loop LLM Quantization Design with Agents

> Source: <https://aiflash.com/news/154098/>
> Published: 2026-11-19 03:00:01+00:00

**Recommendation:** Automate quantization recipe search with an agent loop per model and target precision rather than reusing one hand-tuned recipe across models.
More in the playbook "Inference and self-hosting": <https://wpnews.pro/playbooks/inference-serving>

Post-training quantization is the standard way to serve large language models at low cost. However, the best quantization recipe differs across models and precisions. As a result, experts must redesign it for every new model, which is time-consuming and often sub-optimal. Fortunately, an LLM agent c
