github / github/copilot-cli

/model picker is inconsistent: model step is arrow-keys-only, but the effort/context steps accept direct number input

Open
#3,715 0 comments 0 reactions 0 assignees View on GitHub
area:input-keyboard area:models
Dominant language
Shell
Stars
11.2k
Forks
1.9k
Avg merge
14h 16m
Merged PRs (30d)
6

Description

### Summary
The `/model` flow uses two different interaction patterns across its consecutive steps. Selecting the **model** requires arrow-key navigation, but the **subsequent** prompts (reasoning effort, context, etc.) are numbered and accept direct numeric entry. This forces a hand/keyboard-mode switch mid-flow and is surprising.

### Steps to reproduce
1. Run `/model`.
2. **Step 1 — model list:** all models are listed. The only way to move the selection is the arrow keys; typing `1`, `2`, … does nothing.
3. Pick a model with arrows and press Enter.
4. **Step 2+ — effort / context / etc.:** these options are numbered, so you can immediately press `1` / `2` to choose — no arrows needed.

### Expected
Consistent selection across all steps of `/model`. Either:
- the model list also supports direct numeric quick-select (preferred — faster for long lists), **or**
- all steps use the same single mechanism.

### Actual
Step 1 is arrow-only; later steps are numbered/typeable. The interaction model changes partway through the same command.

### Environment
- Copilot CLI: `1.0.60`
- OS: Windows 11 Pro
- Terminal: WezTerm
- Shell: PowerShell 7.6.2

### Suggested fix
Add numeric quick-select to the model list to match the effort/context pickers (keep arrows working too). This keeps long model lists fast while making the whole `/model` flow consistent.

Contributor guide

Open the contributing guide

Research direction

Start by running the `/model` command and comparing the model-list interaction with the numbered effort and context pickers. Trace the `/model` entry point and the existing picker behavior; done means model choices accept numeric quick-select while arrow navigation still works and the interaction is consistent across all steps.

Written by the indexing model from the issue text.

Assessment

Tech stack
shell
Domain
cli
Issue type
Feature
Difficulty
3/5
Estimated time
1-2 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
64/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.