Google DeepMindLaunchAI ModelsJuly 21, 2026

Google releases Gemini 3.6 Flash and Gemini 3.5 Flash-Lite

Read original source →ai.google.dev

Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are now generally available, both supporting a 1M-token context window, 64k max output tokens, thinking, and Computer Use. The Gemini API also deprecates temperature, top_p, and top_k sampling parameters, which are ignored by these models.

1 source

More stories today

Open the live feed