Evals / image

Colour and tone

2 scenarios, 2 run by an agent. Each run shows its result, time and new tokens.

ScenarioCodexwithout deftCodexwith deft, skill installedClaude Opus 5.5without deftClaude Opus 5.5with deft, skill installedClaude Codewithout deftClaude Codewith deft
Remove the colourMake a black and white version of hopper.png called bw.png.pass9s · 3.0kpass16s · 12.1kpass9s · 8.6kpass13s · 9.5knot runnot run
Brighten a dark photoMake DSCN0010.jpg 20 percent brighter and save it as bright.jpg.pass10s · 5.4kpass13s · 8.4kpass13s · 8.9kpass14s · 9.9knot runnot run