scripod.com
GLM 5.2: why I’m replacing Opus in Claude Code with this new model

Overview

Shownote

Highlights

Transcript

Chapters

Pins

GLM 5.2: why I’m replacing Opus in Claude Code with this new model

How I AI

1 DAYS AGO
How I AI

How I AI

1 DAYS AGO
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

I put GLM 5.2, the open-weight coding model from Z.AI, through four real tasks inside my actual codebase: a codebase architecture audit, a UI redesign, and a 45-minute autonomous bug-hunting session pulling from Sentry and Vercel logs. Total cost: $3.36 fo...

Highlights

This podcast episode features a hands-on evaluation of GLM 5.2, an open-weight coding model from Z.AI, tested across four real-world software development tasks. The host, Claire Vo, assesses the model's performance, cost, and practicality by integrating it into her workflow and running it against her actual codebase.
00:00
Testing GLM 5.2 for Opus-level reasoning at lower cost.
01:39
Opus-level intelligence at a fraction of the price
04:02
Rivals Claude Opus 4.8 and GPT 5.5 on coding tasks.
06:04
Claude Code has better documentation for this process
08:41
Configure Claude Code and Cursor to use GLM 5.2 via OpenRouter
11:04
GLM 5.2 is fast and accurate
12:44
Impressed with the open-weight model's performance.
19:34
Fast, cost-effective, and capable of good design output.
20:58
Model autonomously built a prioritized bug-fix plan
22:36
Excels at HTML/CSS and tool queries
23:50
It excels at autonomous bug-hunting and prioritization.
25:23
GLM 5.2 cost $3.36 for 6 million tokens.

Chapters

What open-weight models are and why GLM 5.2 is worth testing
00:00
GLM 5.2 model overview
01:38
Capabilities and benchmark results
04:02
How to set up GLM 5.2 in Cursor
06:02
How to set up GLM 5.2 in Claude Code
08:37
Live test 1: codebase exploration and architecture audit on ChatPRD
11:04
Live test 2: generating an HTML architecture and roadmap page
12:43
Live test 3: redesigning the How I AI landing page in Cursor
16:37
Live test 4: 45-minute autonomous task, pulling Sentry errors and Vercel logs
20:57
Where it struggled
22:35
My verdict on the output
23:49
Cost breakdown
25:23

Transcript

Claire Vo: What if I told you you could get Opus-level reasoning at a fraction of the cost? That's what we're going to see and test today when I take a look at GLM 5.2. This is our first of many reviews of open-weight and open-source models. To see if we s...