[General] Seeing higher token consumption for similar tasks between the Github Copilot ...

**Mood:** 😔
**Category:** General

Seeing higher token consumption for similar tasks between the Github Copilot CLI and the desktop Github Copilot app, is this expected or a known issue?

---
| Field | Value |
| --- | --- |
| App version | 1.0.13 |
| OS | macOS 26.5.0 |
| Theme | Xotopio |
| Path | /chat |
| Tenure | Week 3 |

Mostly using interactive mode with GPT5.5 with normal context window in High reasoning and default agent. In the CLI I use similar specs.

The report concerns /chat in the desktop app and comparable interactive CLI use; no source file or test is identified. Start by reproducing both flows with GPT5.5, normal context, High reasoning, and the default agent, then compare token accounting; done means the difference is confirmed with evidence and its expected behavior is established.

Vom Indexierungsmodell aus dem Issue-Text verfasst.