ops-dashboard · cc-opus-5.5
claude-code · started 2026-10-08T14:47:57.332Z · horizon 420s · judge ai (openai:gpt-5.5), 5 frames scored
Area under curve
0.930
headline, 0-1
Tab title/icon
12.7s
browser chrome, not the page
First render
12.7s
anything on screen
First reviewable
28.6s
worth feedback
Final score
1.00
no regression
Correctness over time
Where the time went
99%
Model thinking 42.7sTool round trips 0.3s
| Phase | Time | Share | Side |
|---|---|---|---|
| Model thinking | 42.7s | 99.2% | agent |
| Tool round trips | 0.3s | 0.8% | agent |
| Dependency install | 0.0s | 0.0% | toolchain |
| Build | 0.0s | 0.0% | toolchain |
| Dev server boot | 0.0s | 0.0% | toolchain |
| First paint | 0.0s | 0.0% | toolchain |
| Unaccounted | 0.0s | 0.0% | unknown |
| Wall clock (cold start) | 43.0s | 100% |
Agent or toolchain
100%
Agent (model + tool round trips) 43.0s
Toolchain (install + build + boot + paint) 0.0s
Unaccounted 0.0s
Accounted for: 100.0% of wall clock. Agent reported 46.1s of API time; this run attributed 42.7s (delta 3.4s).
- Excludes 15.0s of deliberate idle observation (settle window); the curve still covers it.
Iteration: prompt to visible change
Follow-ups were injected into the same live session.
| Edit | First visible change | Correct change | Agent done | Tool calls | After agent | Broken for |
|---|---|---|---|---|---|---|
| Make the page heading blue. | 3.2s | 3.2s | 6.1s | 1 (Edit) | -2.9s | -- |
| Add an eleventh depot row for Porto: 6,480 shipments, 93.1% on-time, $19.95 cost per shipment. | 3.7s | 3.7s | 8.8s | 1 (Edit) | -5.1s | -- |
Judged frames
Only the header and section labels are visible; required dashboard data is missing.
Visible header and KPIs look polished, but table, bars, and footer are missing from view.
Strong dashboard implementation, but visible screenshot lacks the bar chart and footer.
Strong dashboard matching brief; only the required footer is not visible.
All requested dashboard elements are visible, accurate, and deliberately styled.
Caveats
- judge: openai:gpt-5.5 does not accept a sampling temperature, so these frames were scored at the provider's default rather than at 0. Two runs of the same page can be scored differently, and that variance is inside the AUC. result.json records the temperature as null rather than as the value that was asked for, so p2p compare will not pool this run with one that was scored at a fixed temperature.