PulseAugur
EN
LIVE 16:16:14

Google releases Gemini 3.6 Flash and 3.5 Flash-Lite, deprecating temperature tuning

By PulseAugur Editorial · [1 sources] ·

Google has released two new specialized Gemini models: Gemini 3.6 Flash and Gemini 3.5 Flash-Lite. Gemini 3.6 Flash is an improved, more cost-effective model for general tasks, focusing on reduced output verbosity and better agentic planning. Gemini 3.5 Flash-Lite is designed for high-volume automation and subagent tasks in multi-agent systems. A significant change across both models is the deprecation of temperature, top_p, and top_k sampling parameters, requiring developers to use system instructions for output control and determinism. AI

IMPACT Shifts control from sampling parameters to system instructions, requiring developers to adopt more robust prompt engineering for output determinism.

RANK_REASON New specialized models released by a frontier lab (Google) with significant changes to API behavior. [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on dev.to — LLM tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Google releases Gemini 3.6 Flash and 3.5 Flash-Lite, deprecating temperature tuning

COVERAGE [1]

  1. dev.to — LLM tag TIER_1 English(EN) · albe_sf ·

    Gemini's New Flash Models Change How You Control Outputs

    <p>Google just pushed Gemini 3.6 Flash and 3.5 Flash-Lite to general availability. While the new models target specific builder needs—cost-effective subagents and more efficient planning—the most significant change is the deprecation of <code>temperature</code>, <code>top_p</code…

RELATED ENTITIES

RELATED TOPICS