AT&T reportedly cut AI task costs 56% by routing work to cheaper models
A user says the savings came with a 2% quality loss and contrasts that with Uber spending its entire annual AI budget in four months.
TLDR
A user reports that AT&T cut AI task costs 56% by routing work to cheaper models, with a 2% quality loss, while Uber spent its entire annual AI budget in four months. They argue this favors routing, flat-rate access and owned or open models over using the strongest proprietary model for every request.
The user also summarizes a Mozilla report: open models handled roughly 20% of measured OpenRouter usage but captured only about 4% of model-layer revenue in the cited 2025 window. The summary says Mozilla cautioned that the revenue measurement predates its 2026 usage data, so the current revenue split may differ.
Combined views
—
1 Source, first seen ago
