Claude Opus 5.5 in Antigravity: TwentyDollarAI Standard Benchmarking
Interrupted by the five-hour limit. Usage notes, plus the output.
In each benchmark I run one real task and record what happened: the measured results, where the numbers came from, and the output as evidence.
Interrupted by the five-hour limit. Usage notes, plus the output.
Time, tokens, cost and weekly usage for one run, plus the output.