Milq vs Claude Code
We tested Milq against Claude Code across different categories of apps. All speed tests are done with the same prompt and with Opus 4.6 with High effort. Claude Code was given commands to make changes without approval.
Prompt
“I frequently order food and want to track my delivery without constantly opening the app. Build an iOS app where I can place a mock order and see its progress update in real time—from “Preparing” to “Out for Delivery” to “Arriving Soon.” The app should start a Live Activity as soon as an order is placed, showing key information like the current status, estimated arrival time, and a countdown timer. The Live Activity must appear on the Lock Screen and in the Dynamic Island, and it should update smoothly as the order progresses. Even if real push updates aren’t available, simulate background updates so the activity continues progressing while the app is not in the foreground. The experience should feel responsive, reliable, and native to iOS.”
Milq
Time:04:03:48
- ✓User can place a mock order
- ✓Order transitions through states (Preparing → Out for Delivery → Arriving Soon)
- ✓In-app UI reflects current order state
- ✓Live Activity starts when order is placed
- ✓Displays current status text
- ✓Displays ETA or countdown timer
- ✓Updates over time automatically
- ✓Visible on Lock Screen
- ✓Dynamic Island UI implemented
- ✓Activity continues updating when app is backgrounded (simulation acceptable)
Claude Code
Time:15:23:56
- ✓User can place a mock order
- ✓Order transitions through states (Preparing → Out for Delivery → Arriving Soon)
- ✓In-app UI reflects current order state
- ✓Live Activity starts when order is placed
- ✓Displays current status text
- ✓Displays ETA or countdown timer
- ✓Updates over time automatically
- ✓Visible on Lock Screen
- ✓Dynamic Island UI implemented
- ✓Activity continues updating when app is backgrounded (simulation acceptable)
Build Quality is measured as a weighted average of Compile on First Shot (30%) and Feature Implementation (70%).