Blind Model A/B — "Who's the daily driver?"
2026-07-01 · 5 prompts · Opus 4.8 vs GPT-5.5 · neutral judge: Gemini 3.1 Pro
Two anonymous answers per prompt, re-randomized per prompt so position tells you nothing.
Tap the one you think is stronger on each axis. When you've picked, hit Reveal — it unseals
who wrote what, shows the judge's blind scores, and stacks your picks against the judge's.