[ENHANCEMENT] Update Z.AI Model Descriptions

[bryce-hoehn]

Problem (one or two sentences)

Minor annoyance with Z.AI integration. Basically every model describes itself as the latest model, essentially the descriptions just haven't been updated since.

Example:

"GLM-4.5 is Zhipu's latest featured model. Its comprehensive capabilities in reasoning, coding, and agent reach the state-of-the-art (SOTA) level among open-source models, with a context length of up to 128k."

"GLM-4.6 is Zhipu's newest model with an extended context window of up to 200k tokens, providing enhanced capabilities for processing longer documents and conversations."

"GLM-4.7 is Zhipu's latest model with built-in thinking capabilities enabled by default. It provides enhanced reasoning for complex tasks while maintaining fast response times."

etc.

Context (who is affected and when)

Affects users of Z.ai models, particularly subscribers of the coding plan. It's a bit confusing for users trying to decide which model to use. For the record, as a subscriber to Z.AI Coding Lite, GLM 5.1 works great with Zoo Code and I rarely hit usage limits.

Desired behavior (conceptual, not technical)

Just some minor rewording, like "GLM-4.7 was Zhipu's flagship late-2025 model with built-in thinking capabilities enabled by default. It provides enhanced reasoning for complex tasks while maintaining fast response times."

Not sure where these descriptions are sourced from or if there's a more updated description available.

Also, ideally there would be some indication of the per-model rate limits. From Z.AI:

Below are the current rate limits for each model.

Model type	Model name	Concurrency limit
Language Model	GLM-4.6	3
Language Model	GLM-4.6V-FlashX	3
Language Model	GLM-4.7	2
Image Generation Model	GLM-Image	1
Language Model	GLM-5-Turbo	1
Language Model	GLM-5V-Turbo	1
Language Model	GLM-5.1	10
Language Model	GLM-4.5	10
Language Model	GLM-4.6V	10
Language Model	GLM-4.7-Flash	1
Language Model	GLM-4.7-FlashX	3
Language Model	GLM-OCR	2
Language Model	GLM-5	2
Language Model	GLM-4-Plus	20
Language Model	GLM-4.5V	10
Language Model	GLM-4.6V-Flash	1
Language Model	AutoGLM-Phone-Multilingual	5
Language Model	GLM-4.5-Air	5
Language Model	GLM-4.5-AirX	5
Language Model	GLM-4.5-Flash	2
Language Model	GLM-4-32B-0414-128K	15
Image Generation Model	CogView-4-250304	5
Real-time Audio-Video Model	GLM-ASR-2512	5
Video Generation Model	ViduQ1-text	5
Video Generation Model	Viduq1-Image	5
Video Generation Model	Viduq1-Start-End	5
Video Generation Model	Vidu2-Image	5
Video Generation Model	Vidu2-Start-End	5
Video Generation Model	Vidu2-Reference	5
Video Generation Model	CogVideoX-3	1

Seems like 5.1 is a safe default.

Constraints / preferences (optional)

I'm not sure if these descriptions were written by hand from Roo Code devs or if they sourced the descriptions from somewhere, and if so where. I was going to do a quick PR with updated descriptions from Z.AI but from a quick search I didn't find a good short & sweet description I could copy.

Of course this is a fast moving environment and it's going to be annoying to update constantly. Not sure if there's a better way to automate it?

Request checklist

• I've searched existing Issues and Discussions for duplicates
• This describes a specific problem with clear context and impact

Roo Code Task Links (optional)

No response

Acceptance criteria (optional)

No response

Proposed approach (optional)

No response

Trade-offs / risks (optional)

No response

[bryce-hoehn]

My bad, those rate limits only apply to the API, not the coding plan.

From Z.AI:

GLM-5.1 and GLM-5-Turbo are advanced models designed to rival Claude Opus model. Its usage will be deducted at 3 × during peak hours and 2 × during off-peak hours. We recommend switching to GLM-5.1 for complex tasks and continuing to use GLM-4.7 for routine tasks to avoid rapid quota consumption.

As a limited-time benefit, GLM-5.1 and GLM-5-Turbo will only consume 1× quota during off-peak hours, valid through the end of June.

So GLM-4.7 is probably the safer default.

[roomote]

Closing manually because PR #90 merged into main on 2026-05-18, but the PR body still had the template placeholder instead of a valid closing keyword, so GitHub did not auto-close this issue.