Dung (Donny) Nguyen

Senior Software Engineer

Claude Model Lineup in Copilot

Claude Haiku (e.g., 3.5 Haiku)

Best for: speed and high-volume, simple tasks

Claude Sonnet (3.5 / 3.7 / 4 / 4.5)

Best for: the everyday “workhorse” — balanced speed and capability

This tier is the default choice for most daily development.

Claude Opus (e.g., Opus 4 / 4.1 / 4.8)

Best for: the hardest problems requiring deep reasoning

Opus is the most capable (and most “expensive” in premium-request terms), so it’s typically reserved for tasks where the extra reasoning pays off.

Quick Decision Guide

Need Model tier
Fast completions, simple edits, low cost Haiku
Day-to-day coding, agent tasks, tests, reviews Sonnet
Deep reasoning, complex architecture, tricky debugging Opus

Practical Notes for Copilot

Rule of thumb: Start with Sonnet. Drop to Haiku when you want speed and are doing something simple. Reach for Opus only when Sonnet struggles with a genuinely hard problem.

Note: Anthropic and GitHub update the available model list frequently, so check Copilot → model picker in VS Code for exactly which Claude versions your plan currently offers.