News

Public · Published

Perplexity Fine-Tuned a Chinese AI Model to Match Claude Opus 4.8 at One-Third the Cost

Perplexity fine‑tuned an open‑source Chinese AI model called GLM five point two by adding a high‑level advisor, and the resulting system performs about as well as Claude Opus four point eight while costing roughly one‑third as much.

Published:

Updated:

What happened

Perplexity fine‑tuned an open‑source Chinese AI model called GLM five point two by adding a high‑level advisor, and the resulting system performs about as well as Claude Opus four point eight while costing roughly one‑third as much.

Confirmed

Global impact / market context

Cheaper, high‑performing AI lets more companies in China use advanced language tools, lowering operating costs and encouraging wider adoption, which can boost demand for local AI talent and reduce dependence on expensive foreign models.

Analyst inference

The global AI field is a race to deliver powerful models at lower prices; Perplexity’s cost‑effective approach adds pressure on big providers to improve efficiency, which may lead to price cuts or new pricing strategies for customers.

Analyst inference

What to watch

  1. How quickly Chinese firms adopt GLM five point two, measured by deployment announcements and usage metrics, will show whether the lower cost translates into broader market penetration. Proposed
  2. Reactions from competing AI companies, such as adjusting prices or releasing new models, will indicate how the industry responds to Perplexity’s cheaper high‑performance offering. Proposed
  3. Any new Chinese regulations on AI model fine‑tuning could affect rollout speed, influencing investors’ view of risk and potential returns for firms relying on such technology. Proposed

Affected assets

  • GLM — Golem

Evidence