Shape
What it's good and bad at
What one run costs
The bill, itemised
Tokens used
Everything else
Every single mark
All — checks, and what the graders said
Two AIs graded each of these independently. Where they disagreed we show both — that disagreement is a finding, not an error to hide. Turn on Nerd mode for the graders' written reasoning.
| ID | Question | Skill | Score | What the graders said |
|---|