Gemini last models: temperature, top_p, and top_k are deprecated and ignored

Google has released Gemini 3.6 Flash and Gemini 3.5 Flash-Lite models, with the former offering stronger performance on complex tasks at lower costs than the 3.5 Flash, and the latter providing the fastest, cheapest option in the 3.5 family. The update notes that temperature, top_p, and top_k parameters are deprecated and ignored in the latest Gemini models. Both new models support a 1M token context window, 64k max output tokens, and full built-in tools including Computer Use.