AI Scorecard

Weekly provider comparison from real-world reviews.

Every week, AntFleet publishes structured comparison data from real-world reviews on opted-in customer code. Two frontier models (Claude Opus 4 + GPT-5) run independently; only the findings both agree on get posted. The numbers below show what each model produces over time. This is rolling data — not a static benchmark, not a marketing exercise. Methodology.