Gemini 3.5 Flash-Lite tops out at nearly 350 tokens per second

OfficialLoganK · x · 2026-07-21

Google’s Gemini 3.5 Flash-Lite is being positioned as the company’s smallest and fastest Gemini model.

According to the post, it is:

The poster says the latency makes it feel especially smooth for UI experiences and even viable for agent harnesses.

Related event: Google Launches Three New Gemini Models, Announces Gemini 4 Pre-training(157 posts)→

Original post →

More from coding & agent

coding & agent channel →