GitHub Copilot CLI adds GPT-5.5

Hot take: this is a practical model upgrade, not just a checkbox release. The real value is in pairing a stronger model with CLI-native agent modes that can sustain longer, messier coding tasks.

• –GPT-5.5 in the model picker makes Copilot CLI more credible for deep refactors, debugging, and multi-step implementation work.
• –Plan mode and autopilot are the differentiators here: the workflow is about staying in the terminal while the agent carries out the sequence.
• –This is a product-update story more than a model-release story, because the user-facing change is how the CLI behaves in real coding sessions.
• –The update positions Copilot CLI more directly against other terminal agents by improving both model choice and workflow continuity.