Definition
Gemini Flash is Google’s efficiency-tier gemini model family optimized for latency, token cost, and high-throughput agentic workloads versus Pro/Ultra frontier tiers.
Key Points
- 2026-07-21: gemini-36-flash replaces 3.5 Flash as workhorse (~17% fewer output tokens; 7.50)
- gemini-35-flash-lite for ultra-cheap high-throughput (~2.50, ~350 tok/s)
- Flash base also used for specialized gemini-35-flash-cyber