Hands-on: Gemini 3.6 Flash is Slightly More Token-Efficient Than 3.5

scaling01 · x · 2026-07-21

The author notes that Gemini 3.6 Flash is slightly more token-efficient than version 3.5. However, they argue that Grok 4.5 or Sol remain the top choices for lower reasoning settings right now.

Related event: Gemini 3.6 Flash Benchmarks Leak: Stagnant Performance but Doubled Efficiency(18 posts)→

Original post →

More from Models

Models channel →