AI on the Clock Türkçe mehmeterkek.com

Pulse

15 August 2026

1 story, chosen and edited by hand.

01

Introducing Gemini 3.7 Flash

Google DeepMind released Gemini 3.7 Flash three weeks after 3.6 Flash, at an introductory price of $0.75/1M input and $3.75/1M output tokens through the end of the year — half the 3.6 cost. Reported gains: DeepSWE v1.1 65.3% vs 49.0%, FrontierCode 1.1 Main 43.6% vs 34.4%, GDP.pdf 34.0% vs 22.0%, WebDev Arena Elo 1588 vs 1538. Gemini Spark switches to it today.

My readA three-week gap and DeepSWE jumps 16 points while the price halves. That's the part that got me — not the intelligence claim, the cost curve. But read AutomationBench honestly: 30.4% up from 17.0% means it still blows nearly seven of ten real business workflows. Great for cheap retries, not yet for unattended agents. I'll be watching what happens to that $0.75 in January.

Read the original at Google DeepMind →