[CLI] GLM-5.3 Flash reasoning effort is grayed out / cannot be changed

[myplxdm]

What problem are you trying to solve?

Summary

On Freebuff CLI(macOS), the reasoning/thinking effort control for GLM 5.3 Flash is grayed out and cannot be changed. The model always runs at deep reasoning (max), which causes very long thinking time and repeated-looking thought streams in agent workflows.

Environment

• Product: Freebuff CLI
• OS: macOS (Apple Silicon / Intel — please fill)
• App version: 0.0.162
• Model: GLM 5.3 Flash (z-ai/glm-5.3-flash)

Steps to reproduce

1. Open Freebuff Desktop and select GLM 5.3 Flash
2. Open the model / reasoning effort settings
3. Observe that the effort control is grayed out and cannot be set to low or high

Expected

GLM-5.3 Flash officially supports reasoning_effort: low | high | max (default max). Users should be able to select low or high for faster responses on simpler tasks, same as other models that expose an effort ladder.

Actual

• Effort UI is disabled/grayed out
• Requests appear to always use the deepest reasoning setting
• Long thinking time; thinking stream often looks repetitive, but the task can still complete eventually

Additional context

• Same limitation on CLI (no way to set effort)
• Official Z.AI docs: thinking cannot be disabled on GLM-5.3 / GLM-5.3-Flash, but effort levels (low/high/max) are supported
• Freebuff model catalog already defines efforts for this model; UI does not expose a working control

Request

Please enable the reasoning effort picker for GLM 5.3 Flash on Desktop (and ideally CLI), or document that effort is intentionally fixed to max for this model.

What would you like to happen?

Enable the reasoning effort control for GLM 5.3 Flash on Desktop (and CLI if possible), so users can choose low / high / max instead of being stuck on max with a grayed-out control.

Area

CLI (terminal client)

Contribution

• I would be willing to open a PR for this