# @qwweryo — 2026-08-29

♥2 ↻0 · https://x.com/qwweryo/status/2093752209489973524

GLM-5.3-Flash is the hot new open source model (Ox Alpha) and seems to be better than Opus 4.8

Unsloth shipped a 3-bit quantized version that needs 128GB of RAM just for the weights

So let's say you have 5 RTX 5090s for $20k

You'll likely get very slow token output

tags: author:qwweryo, kind:tweet, model:claude-opus-4-8, on:glm-5-3-flash, year:2026
cited on: _dossiers/glm-5-3-flash.md, glm-5-3-flash