BLAST RADIUS

Open-weight & owned stack · Training tuning

Rogueon previews GLM-based DeepThink reasoning model

Sep 24, 2026

Rogueon previewed DeepThink, a reasoning model built on GLM, without documenting a broader public release or deployment.

Read the original at reddit.comOpens the publisher's site in a new tab

More in Open-weight & owned stack