Claude Opus 5.5 in Antigravity: TwentyDollarAI Standard Benchmarking
Interrupted by the five-hour limit. Usage notes, plus the output.
Results
Time is my estimate. Quota from my Google Pro usage meter.
| Time (total runtime) | About 24m |
|---|---|
| Tokens (total) | Not measured |
| Cost (API-equivalent) | Not measured |
| Five-hour limit | Hit mid-test |
| Weekly usage | Shared (see notes) |
Models tested
- Claude Opus 5.5Plan: Google Pro plan
The prompt
Standard TwentyDollarAI Benchmarking
Evidence
My take
- Attractive scene: recognizable Maltese, rich meadow.
- Weaker fidelity: toy and dog less true to the reference.
- Limits matter: testing paused. Scores and notes
I post new tests on X: @TwentyDollarAI (opens in a new tab)
← All benchmarks