Google Slashes AI Costs

Instead of chasing raw intelligence, Google just quietly pivoted to aggressive cost-cutting with the surprise release of Gemini 3.6 Flash on July 21st, 2026. While users were still waiting for the delayed Gemini 3.5 Pro, this mid-cycle update bypassed flashy frontier claims to deliver a decisive 17% efficiency gain.

It matches its predecessor's performance on applied benchmarks like coding and computer use, yet it flatlines at a score of 50 on the Artificial Analysis Intelligence Index. This memorable shift is all about economics, dropping output token costs from $9 to $7.50 per million.

By slashing high-volume operational costs alongside a new Flash Lite release, Google is proving that the next phase of the AI race might be one on sheer affordability.